fix(tools,gitea): alias repo/index args and reject empty path segments

Per-repo MCP tools 404'd while owner-level tools worked (#36). Root cause was
a parameter-name contract mismatch, not a routing fault: every per-repo tool
declares the repo identifier as 'name' (and issue/PR index as 'number'), but
every caller — the claude.ai connector, LLMs primed on gitea's own API — sends
'repo' and 'index'. The unmatched fields zero-valued the upstream path segment,
producing '/api/v1/repos/{owner}//...', which gitea answers with its generic
api-404 whose body points at /api/swagger. That swagger pointer is gitea boiler-
plate, not a misroute — the MCP dispatch was correct all along.

Two layers of defence:
- parseArgs aliases repo->name and index->number (explicit canonical wins;
  alias key left intact so pr_merge's real 'index' field is unaffected). Kills
  the recurrence by accepting the idiomatic argument names.
- the gitea client rejects any path with an empty segment before the HTTP call,
  returning ErrValidation instead of forwarding a malformed path and surfacing
  gitea's opaque swagger-404. Kills the silent-misleading-error class.

Closes #36

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

.aider.conventions.md

@@ -36,6 +36,18 @@ These rules apply to every task across every project, regardless of harness.
 4. **Goal-driven execution.** Define clear success criteria up front for every task.
    Loop — implement, verify, refine — until those criteria are met. Don't claim
    completion without evidence (tests pass, command output, observed behavior).
+5. **Trunk-Based Development — commit directly to main.** Every commit is one
+   logical change (one tool, one fix, one test) with passing tests. Main is always
+   deployable. Never create long-lived feature branches.
+
+   **Exception — parallel agents on same repo:** If another agent is known to be
+   actively working on the same repo simultaneously, create a short-lived branch
+   (`agent/<description>`), finish the task, and merge to main within the same
+   session. Do not leave agent branches open between sessions.
+
+   **Exception — external contributor or client four-eyes requirement:** Use
+   PR flow only when a human reviewer outside the project is required. Document
+   the reason in PROJECT.md.
 
 ## Default stack
 
@@ -49,6 +61,7 @@ These rules apply to every task across every project, regardless of harness.
 | Search | pgvector (vector), BM25 | Qdrant (when >1M vectors or hybrid retrieval) | — |
 | Logging | slog (structured) | — | — |
 | Testing | Table-driven, testify | — | — |
+| Agents (Go) | google.golang.org/adk + pkg/litellm adapter | — | — |
 
 Exploratory: Rust, Zig — I'll tell you when I want these.
 
@@ -58,9 +71,12 @@ Exploratory: Rust, Zig — I'll tell you when I want these.
 - **Errors**: `fmt.Errorf("operation: %w", err)` — never naked, never log-and-return
 - **Naming**: stdlib conventions, no stuttering
 - **Architecture**: prefer stdlib over frameworks, constructor injection, env-var config parsed into typed structs
-- **Git**: conventional commits (`feat:`, `fix:`, `chore:`), one concern per PR, PR describes *why* not *what*
+- **Git**: conventional commits (`feat:`, `fix:`, `chore:`), commit directly to main,
+  one logical change per commit, CI is the quality gate
+- **Never**: long-lived feature branches, PRs for solo work, direct push without
+  passing `task check` locally first
 - **Security**: no secrets in code, govulncheck before adding deps, SOPS for encrypted config
-- **Dependencies**: prefer stdlib. testify, slog, templ, sqlc are pre-approved; anything else needs justification in the commit message
+- **Dependencies**: prefer stdlib. testify, slog, templ, sqlc, google.golang.org/adk (agent projects only) are pre-approved; anything else needs justification in the commit message
 
 ## Infrastructure
 
@@ -100,18 +116,64 @@ See `~/dev/PROJECT_SUMMARY.md` for detailed descriptions of each project.
 - **koala-ai-stack** (`AGENTS/`) — local AI server infrastructure management
 - **klimatkollen** (`XT/`) — Swedish municipal climate data platform
 
-## Knowledge base
+## Knowledge base — actively use it
 
-When available, agents can query the shared knowledge base:
+A persistent brain (BM25 search + LLM-synthesised Q&A) survives across sessions,
+hosts, and harnesses. It holds 100+ hard-won entries: infra incident postmortems,
+Go pitfalls, framework gotchas, design principles, ADRs. **It is not optional
+reference material — query it actively, not just when explicitly told.**
 
-- **MCP**: `mcp://hyperguild.<TAILNET>.ts.net:3100/knowledge`
-- **HTTP**: `http://hyperguild.<TAILNET>.ts.net:3100/api/v1/search`
+### When to query (treat as a reflex)
 
-<!-- TODO: replace <TAILNET> placeholder with the real Tailscale tailnet
-     name once hyperguild is deployed. Until then, agents that try to
-     reach the knowledge service on a host where it isn't running will
-     get DNS NXDOMAIN, which is the desired fail-loudly behavior. -->
-- **Scoping**: defaults to `public` collection; client projects filter to `{client}` + `public`
+- **Before** starting a non-trivial task — search for prior art with the symptom
+  AND the system component ("how did we solve X in Y?"). 5 seconds beats 5 hours.
+- **When debugging** — search for the error string, the stack frame, the affected
+  service. Past you may have already paid this tax.
+- **Before adopting** a pattern, library, framework, or model name — check if it
+  was tried and rejected, or what the integration footguns are.
+- **When making architectural decisions** — search for the domain + "ADR" or
+  "decision" to find prior reasoning before re-deriving it.
+- **When a recommendation feels novel** — challenge yourself: "has this been
+  documented?" The brain often has it.
+
+### When to write
+
+After you discover something that **future-you would forget** and that **isn't
+recoverable from the code, git log, or PR description alone**:
+
+- Bugs whose root cause is non-obvious and generalisable beyond this project.
+- Framework / library / model-name quirks that bit you and would bite anyone.
+- Design principles validated under fire (e.g. "every `_get` needs a `_list`").
+- Postmortems for incidents: what broke, why, how diagnosed, what to do next time.
+
+DON'T write project status, sprint progress, PR summaries, or "what I did this
+session" — those rot fast and the originals are in git/gitea anyway. Brain
+entries that age well are about *why*, *how to avoid*, and *what to do when*.
+
+### How to access (per harness)
+
+| Harness | Query | Write |
+|---------|-------|-------|
+| **Claude Code, Claude Desktop** | `brain_query` (BM25), `brain_answer` (LLM-synth + sources) MCP tools | `brain_write` MCP tool |
+| **Crush, Pi, Antigravity, other MCP-capable** | same MCP server: `ingestion-brain` (via the `mcp__*_brain__*` namespace once authenticated) | same |
+| **Anything HTTP-only (curl, scripts)** | `POST https://brain-mcp.d-ma.be/query` with `{"query":"..."}` (auth via `BRAIN_MCP_TOKEN`) | `POST .../write` with `{"content":"...","filename":"..."}` |
+| **Browser / human inspection** | `https://gitea.d-ma.be/mathias/hyperguild` → `knowledge/` and `wiki/` markdown files |
+
+- **Scoping**: defaults to `public` collection; client projects filter to `{client}` + `public`.
+- **Routing**: brain_answer's LLM uses berget.ai as primary, iguana ollama as
+  fallback. Both are configurable in the `supervisor/ingestion-deployment.yaml`
+  on the koala k3s cluster; don't hardcode local-only model names into the
+  berget URL (see knowledge entry on namespace mismatches).
+
+### Quick reflex checks
+
+If you find yourself about to say any of these out loud, you owe yourself a brain query first:
+
+- "I think the issue might be..."
+- "Let me try X and see..."
+- "I'll just write a script to..."
+- "This is probably a new bug..."
+- "Has anyone done this before?" — *yes, probably, go check.*
 
 ## Client work rules
 
@@ -208,9 +270,11 @@ Key skills:
 
 ### Git
 - Conventional commits: `feat:`, `fix:`, `chore:`, `docs:`, `refactor:`
-- Branch naming: `feat/short-description`, `fix/short-description`
-- PRs: one concern per PR, description explains *why* not *what*
-- **Branch protection:** always work on a feature branch, open a PR, never push directly to main
+- **Trunk-Based Development:** commit directly to main. One logical change per commit.
+- Run `task check` locally before every push. CI is the quality gate, not branch protection.
+- No feature branches, no PRs for solo/agent work.
+- Exception: if a parallel agent session is active on this repo, use a short-lived
+  `agent/<description>` branch and merge within the same session.
 
 ### Security
 - No secrets in code, ever — use env vars or SOPS-encrypted files
@@ -248,98 +312,29 @@ When acting as a coding agent on this project:
 3. If unsure about a convention, check `DECISIONS.md` or ask
 4. Never modify files outside the project root without explicit permission
 5. When adding a dependency, explain why in the commit message
-6. Always work on a feature branch and open a PR — never push directly to main
+6. Commit directly to main. Run `task check` before every push. Never create
+   feature branches unless a parallel agent is simultaneously active on this repo.
 7. For client projects: never send code or context to cloud APIs — use local models via LiteLLM
 
-## Current sprint — gitea-mcp v0.2 patch (2026-05-14)
+## Current state — v0.2.5 (2026-05-17)
 
-### Context
-The main v0.2 batch (repo_create, repo_update, repo_mirror_push, repo_delete,
-repo_tree, repo_topics_update, file_read dir-fix, issue_get, release_create,
-create_project_from_template) was implemented and pushed directly to main.
+All v0.2 work is complete and deployed. No active sprint.
 
-This sprint fixes three remaining gaps found during code review on 2026-05-14.
-These are blockers for `hyperguild new-project`.
+### What shipped
 
-### Issues to fix (all three in one PR: `fix/v02-patch`)
+| Tag | PR | Tools / fixes |
+|-----|----|---------------|
+| v0.2.2 | #21 | repo_create, repo_update, repo_mirror_push |
+| v0.2.3 | #22 | repo_tree, repo_topics_update, file_read dir fix |
+| v0.2.4 | #23 | issue_get, release_create, repo_delete |
+| v0.2.5 | #26 | repo_update archived+template, create_project_from_template template_name, pr_files_diff loop fix |
 
-#### #12 — repo_update: add `archived` and `template` fields
-**File:** `internal/gitea/repos.go` → `UpdateRepoArgs` struct
-**File:** `internal/tools/repo_update.go` → input schema + args struct
+Current main: `e31fd3f`. CI green. Deployed via Flux.
 
-Add to `UpdateRepoArgs`:
-```go
-Archived *bool
-Template *bool
-```
+### Next up
 
-Add to tool input schema:
-```json
-"archived": {
-  "type": "boolean",
-  "description": "Mark repo as archived (read-only). Requires confirm=<repo name>."
-},
-"template": {
-  "type": "boolean",
-  "description": "Toggle template repo flag."
-}
-```
+1. **`hyperguild new-project` v1** — primary next target.
+   See brain node `adr-new-project-gitea-first-github-mirror` for full flow spec.
 
-Add confirm-guard for `archived=true` (same pattern as `private=false`):
-```go
-if args.Archived != nil && *args.Archived {
-    if args.Confirm != args.Name {
-        return nil, fmt.Errorf("setting archived=true is irreversible: set confirm=%q to proceed", args.Name)
-    }
-}
-```
-
-New test cases to add in `repo_update_test.go`:
-- `TestRepoUpdateTool_Archive` — happy path with confirm
-- `TestRepoUpdateTool_ArchiveRequiresConfirm` — missing confirm returns error
-- `TestRepoUpdateTool_SetTemplate` — no confirm needed
-
-#### #24 — create_project_from_template: make template selectable
-**File:** `internal/tools/create_project_from_template.go`
-
-Add optional `template_name` param to input schema:
-```json
-"template_name": {
-  "type": "string",
-  "enum": ["template-go-web", "template-go-agent"],
-  "description": "Template repo to generate from. Defaults to template-go-web.",
-  "default": "template-go-web"
-}
-```
-
-The tool should use `args.TemplateName` if set, fall back to the hardcoded default.
-Remove the hardcoded template name from `cmd/gitea-mcp/main.go` constructor call —
-the tool resolves it internally.
-
-New test case: `TestCreateProjectFromTemplate_AgentTemplate`
-
-#### #25 — pr_files_diff: fix same diff returned for all files
-**File:** `internal/tools/pr_files_diff.go`
-
-There is a loop bug where all file entries in the response contain the same diff
-(the first file's diff is reused for every subsequent file). Find the loop and
-ensure each iteration reads and assigns the correct diff for its own file.
-
-Reproduce: call `pr_files_diff` on any PR with 3+ files, verify each file has
-a distinct diff.
-
-### Definition of done
-
-- [ ] `task check` passes
-- [ ] `repo_update` accepts `archived` and `template` params
-- [ ] `archived=true` requires `confirm=<repo name>`
-- [ ] `create_project_from_template` accepts `template_name` param, defaults to `template-go-web`
-- [ ] `pr_files_diff` returns distinct diff per file
-- [ ] All new test cases pass
-- [ ] PR `fix/v02-patch` merged to main via PR (not direct push)
-
-### After this sprint
-
-Next: `hyperguild new-project` v1 implementation.
-See brain node `adr-new-project-gitea-first-github-mirror` for the full flow spec.
-Also: verify end-to-end mirror flow (issue #19) once `repo_mirror_push` is confirmed working.
+2. **Issue #19** — end-to-end mirror flow verification.
+   `repo_mirror_push` is implemented but the full flow (create repo → add push mirror → verify sync to GitHub) has not been tested manually. Do this before relying on it in production.

.context/PROJECT.md

@@ -37,9 +37,11 @@
 
 ### Git
 - Conventional commits: `feat:`, `fix:`, `chore:`, `docs:`, `refactor:`
-- Branch naming: `feat/short-description`, `fix/short-description`
-- PRs: one concern per PR, description explains *why* not *what*
-- **Branch protection:** always work on a feature branch, open a PR, never push directly to main
+- **Trunk-Based Development:** commit directly to main. One logical change per commit.
+- Run `task check` locally before every push. CI is the quality gate, not branch protection.
+- No feature branches, no PRs for solo/agent work.
+- Exception: if a parallel agent session is active on this repo, use a short-lived
+  `agent/<description>` branch and merge within the same session.
 
 ### Security
 - No secrets in code, ever — use env vars or SOPS-encrypted files
@@ -77,98 +79,29 @@ When acting as a coding agent on this project:
 3. If unsure about a convention, check `DECISIONS.md` or ask
 4. Never modify files outside the project root without explicit permission
 5. When adding a dependency, explain why in the commit message
-6. Always work on a feature branch and open a PR — never push directly to main
+6. Commit directly to main. Run `task check` before every push. Never create
+   feature branches unless a parallel agent is simultaneously active on this repo.
 7. For client projects: never send code or context to cloud APIs — use local models via LiteLLM
 
-## Current sprint — gitea-mcp v0.2 patch (2026-05-14)
+## Current state — v0.2.5 (2026-05-17)
 
-### Context
-The main v0.2 batch (repo_create, repo_update, repo_mirror_push, repo_delete,
-repo_tree, repo_topics_update, file_read dir-fix, issue_get, release_create,
-create_project_from_template) was implemented and pushed directly to main.
+All v0.2 work is complete and deployed. No active sprint.
 
-This sprint fixes three remaining gaps found during code review on 2026-05-14.
-These are blockers for `hyperguild new-project`.
+### What shipped
 
-### Issues to fix (all three in one PR: `fix/v02-patch`)
+| Tag | PR | Tools / fixes |
+|-----|----|---------------|
+| v0.2.2 | #21 | repo_create, repo_update, repo_mirror_push |
+| v0.2.3 | #22 | repo_tree, repo_topics_update, file_read dir fix |
+| v0.2.4 | #23 | issue_get, release_create, repo_delete |
+| v0.2.5 | #26 | repo_update archived+template, create_project_from_template template_name, pr_files_diff loop fix |
 
-#### #12 — repo_update: add `archived` and `template` fields
-**File:** `internal/gitea/repos.go` → `UpdateRepoArgs` struct
-**File:** `internal/tools/repo_update.go` → input schema + args struct
+Current main: `e31fd3f`. CI green. Deployed via Flux.
 
-Add to `UpdateRepoArgs`:
-```go
-Archived *bool
-Template *bool
-```
+### Next up
 
-Add to tool input schema:
-```json
-"archived": {
-  "type": "boolean",
-  "description": "Mark repo as archived (read-only). Requires confirm=<repo name>."
-},
-"template": {
-  "type": "boolean",
-  "description": "Toggle template repo flag."
-}
-```
+1. **`hyperguild new-project` v1** — primary next target.
+   See brain node `adr-new-project-gitea-first-github-mirror` for full flow spec.
 
-Add confirm-guard for `archived=true` (same pattern as `private=false`):
-```go
-if args.Archived != nil && *args.Archived {
-    if args.Confirm != args.Name {
-        return nil, fmt.Errorf("setting archived=true is irreversible: set confirm=%q to proceed", args.Name)
-    }
-}
-```
-
-New test cases to add in `repo_update_test.go`:
-- `TestRepoUpdateTool_Archive` — happy path with confirm
-- `TestRepoUpdateTool_ArchiveRequiresConfirm` — missing confirm returns error
-- `TestRepoUpdateTool_SetTemplate` — no confirm needed
-
-#### #24 — create_project_from_template: make template selectable
-**File:** `internal/tools/create_project_from_template.go`
-
-Add optional `template_name` param to input schema:
-```json
-"template_name": {
-  "type": "string",
-  "enum": ["template-go-web", "template-go-agent"],
-  "description": "Template repo to generate from. Defaults to template-go-web.",
-  "default": "template-go-web"
-}
-```
-
-The tool should use `args.TemplateName` if set, fall back to the hardcoded default.
-Remove the hardcoded template name from `cmd/gitea-mcp/main.go` constructor call —
-the tool resolves it internally.
-
-New test case: `TestCreateProjectFromTemplate_AgentTemplate`
-
-#### #25 — pr_files_diff: fix same diff returned for all files
-**File:** `internal/tools/pr_files_diff.go`
-
-There is a loop bug where all file entries in the response contain the same diff
-(the first file's diff is reused for every subsequent file). Find the loop and
-ensure each iteration reads and assigns the correct diff for its own file.
-
-Reproduce: call `pr_files_diff` on any PR with 3+ files, verify each file has
-a distinct diff.
-
-### Definition of done
-
-- [ ] `task check` passes
-- [ ] `repo_update` accepts `archived` and `template` params
-- [ ] `archived=true` requires `confirm=<repo name>`
-- [ ] `create_project_from_template` accepts `template_name` param, defaults to `template-go-web`
-- [ ] `pr_files_diff` returns distinct diff per file
-- [ ] All new test cases pass
-- [ ] PR `fix/v02-patch` merged to main via PR (not direct push)
-
-### After this sprint
-
-Next: `hyperguild new-project` v1 implementation.
-See brain node `adr-new-project-gitea-first-github-mirror` for the full flow spec.
-Also: verify end-to-end mirror flow (issue #19) once `repo_mirror_push` is confirmed working.
+2. **Issue #19** — end-to-end mirror flow verification.
+   `repo_mirror_push` is implemented but the full flow (create repo → add push mirror → verify sync to GitHub) has not been tested manually. Do this before relying on it in production.

.context/mcp.json

@@ -16,7 +16,10 @@
     },
     "infra": {
       "type": "http",
-      "url": "https://infra-mcp.d-ma.be/mcp"
+      "url": "https://infra-mcp.d-ma.be/mcp",
+      "headers": {
+        "Authorization": "Bearer ${INFRA_MCP_TOKEN}"
+      }
     }
   }
 }

.context/system-prompt.txt

@@ -41,6 +41,18 @@ These rules apply to every task across every project, regardless of harness.
 4. **Goal-driven execution.** Define clear success criteria up front for every task.
    Loop — implement, verify, refine — until those criteria are met. Don't claim
    completion without evidence (tests pass, command output, observed behavior).
+5. **Trunk-Based Development — commit directly to main.** Every commit is one
+   logical change (one tool, one fix, one test) with passing tests. Main is always
+   deployable. Never create long-lived feature branches.
+
+   **Exception — parallel agents on same repo:** If another agent is known to be
+   actively working on the same repo simultaneously, create a short-lived branch
+   (`agent/<description>`), finish the task, and merge to main within the same
+   session. Do not leave agent branches open between sessions.
+
+   **Exception — external contributor or client four-eyes requirement:** Use
+   PR flow only when a human reviewer outside the project is required. Document
+   the reason in PROJECT.md.
 
 ## Default stack
 
@@ -54,6 +66,7 @@ These rules apply to every task across every project, regardless of harness.
 | Search | pgvector (vector), BM25 | Qdrant (when >1M vectors or hybrid retrieval) | — |
 | Logging | slog (structured) | — | — |
 | Testing | Table-driven, testify | — | — |
+| Agents (Go) | google.golang.org/adk + pkg/litellm adapter | — | — |
 
 Exploratory: Rust, Zig — I'll tell you when I want these.
 
@@ -63,9 +76,12 @@ Exploratory: Rust, Zig — I'll tell you when I want these.
 - **Errors**: `fmt.Errorf("operation: %w", err)` — never naked, never log-and-return
 - **Naming**: stdlib conventions, no stuttering
 - **Architecture**: prefer stdlib over frameworks, constructor injection, env-var config parsed into typed structs
-- **Git**: conventional commits (`feat:`, `fix:`, `chore:`), one concern per PR, PR describes *why* not *what*
+- **Git**: conventional commits (`feat:`, `fix:`, `chore:`), commit directly to main,
+  one logical change per commit, CI is the quality gate
+- **Never**: long-lived feature branches, PRs for solo work, direct push without
+  passing `task check` locally first
 - **Security**: no secrets in code, govulncheck before adding deps, SOPS for encrypted config
-- **Dependencies**: prefer stdlib. testify, slog, templ, sqlc are pre-approved; anything else needs justification in the commit message
+- **Dependencies**: prefer stdlib. testify, slog, templ, sqlc, google.golang.org/adk (agent projects only) are pre-approved; anything else needs justification in the commit message
 
 ## Infrastructure
 
@@ -105,18 +121,64 @@ See `~/dev/PROJECT_SUMMARY.md` for detailed descriptions of each project.
 - **koala-ai-stack** (`AGENTS/`) — local AI server infrastructure management
 - **klimatkollen** (`XT/`) — Swedish municipal climate data platform
 
-## Knowledge base
+## Knowledge base — actively use it
 
-When available, agents can query the shared knowledge base:
+A persistent brain (BM25 search + LLM-synthesised Q&A) survives across sessions,
+hosts, and harnesses. It holds 100+ hard-won entries: infra incident postmortems,
+Go pitfalls, framework gotchas, design principles, ADRs. **It is not optional
+reference material — query it actively, not just when explicitly told.**
 
-- **MCP**: `mcp://hyperguild.<TAILNET>.ts.net:3100/knowledge`
-- **HTTP**: `http://hyperguild.<TAILNET>.ts.net:3100/api/v1/search`
+### When to query (treat as a reflex)
 
-<!-- TODO: replace <TAILNET> placeholder with the real Tailscale tailnet
-     name once hyperguild is deployed. Until then, agents that try to
-     reach the knowledge service on a host where it isn't running will
-     get DNS NXDOMAIN, which is the desired fail-loudly behavior. -->
-- **Scoping**: defaults to `public` collection; client projects filter to `{client}` + `public`
+- **Before** starting a non-trivial task — search for prior art with the symptom
+  AND the system component ("how did we solve X in Y?"). 5 seconds beats 5 hours.
+- **When debugging** — search for the error string, the stack frame, the affected
+  service. Past you may have already paid this tax.
+- **Before adopting** a pattern, library, framework, or model name — check if it
+  was tried and rejected, or what the integration footguns are.
+- **When making architectural decisions** — search for the domain + "ADR" or
+  "decision" to find prior reasoning before re-deriving it.
+- **When a recommendation feels novel** — challenge yourself: "has this been
+  documented?" The brain often has it.
+
+### When to write
+
+After you discover something that **future-you would forget** and that **isn't
+recoverable from the code, git log, or PR description alone**:
+
+- Bugs whose root cause is non-obvious and generalisable beyond this project.
+- Framework / library / model-name quirks that bit you and would bite anyone.
+- Design principles validated under fire (e.g. "every `_get` needs a `_list`").
+- Postmortems for incidents: what broke, why, how diagnosed, what to do next time.
+
+DON'T write project status, sprint progress, PR summaries, or "what I did this
+session" — those rot fast and the originals are in git/gitea anyway. Brain
+entries that age well are about *why*, *how to avoid*, and *what to do when*.
+
+### How to access (per harness)
+
+| Harness | Query | Write |
+|---------|-------|-------|
+| **Claude Code, Claude Desktop** | `brain_query` (BM25), `brain_answer` (LLM-synth + sources) MCP tools | `brain_write` MCP tool |
+| **Crush, Pi, Antigravity, other MCP-capable** | same MCP server: `ingestion-brain` (via the `mcp__*_brain__*` namespace once authenticated) | same |
+| **Anything HTTP-only (curl, scripts)** | `POST https://brain-mcp.d-ma.be/query` with `{"query":"..."}` (auth via `BRAIN_MCP_TOKEN`) | `POST .../write` with `{"content":"...","filename":"..."}` |
+| **Browser / human inspection** | `https://gitea.d-ma.be/mathias/hyperguild` → `knowledge/` and `wiki/` markdown files |
+
+- **Scoping**: defaults to `public` collection; client projects filter to `{client}` + `public`.
+- **Routing**: brain_answer's LLM uses berget.ai as primary, iguana ollama as
+  fallback. Both are configurable in the `supervisor/ingestion-deployment.yaml`
+  on the koala k3s cluster; don't hardcode local-only model names into the
+  berget URL (see knowledge entry on namespace mismatches).
+
+### Quick reflex checks
+
+If you find yourself about to say any of these out loud, you owe yourself a brain query first:
+
+- "I think the issue might be..."
+- "Let me try X and see..."
+- "I'll just write a script to..."
+- "This is probably a new bug..."
+- "Has anyone done this before?" — *yes, probably, go check.*
 
 ## Client work rules
 
@@ -213,9 +275,11 @@ Key skills:
 
 ### Git
 - Conventional commits: `feat:`, `fix:`, `chore:`, `docs:`, `refactor:`
-- Branch naming: `feat/short-description`, `fix/short-description`
-- PRs: one concern per PR, description explains *why* not *what*
-- **Branch protection:** always work on a feature branch, open a PR, never push directly to main
+- **Trunk-Based Development:** commit directly to main. One logical change per commit.
+- Run `task check` locally before every push. CI is the quality gate, not branch protection.
+- No feature branches, no PRs for solo/agent work.
+- Exception: if a parallel agent session is active on this repo, use a short-lived
+  `agent/<description>` branch and merge within the same session.
 
 ### Security
 - No secrets in code, ever — use env vars or SOPS-encrypted files
@@ -253,100 +317,31 @@ When acting as a coding agent on this project:
 3. If unsure about a convention, check `DECISIONS.md` or ask
 4. Never modify files outside the project root without explicit permission
 5. When adding a dependency, explain why in the commit message
-6. Always work on a feature branch and open a PR — never push directly to main
+6. Commit directly to main. Run `task check` before every push. Never create
+   feature branches unless a parallel agent is simultaneously active on this repo.
 7. For client projects: never send code or context to cloud APIs — use local models via LiteLLM
 
-## Current sprint — gitea-mcp v0.2 patch (2026-05-14)
+## Current state — v0.2.5 (2026-05-17)
 
-### Context
-The main v0.2 batch (repo_create, repo_update, repo_mirror_push, repo_delete,
-repo_tree, repo_topics_update, file_read dir-fix, issue_get, release_create,
-create_project_from_template) was implemented and pushed directly to main.
+All v0.2 work is complete and deployed. No active sprint.
 
-This sprint fixes three remaining gaps found during code review on 2026-05-14.
-These are blockers for `hyperguild new-project`.
+### What shipped
 
-### Issues to fix (all three in one PR: `fix/v02-patch`)
+| Tag | PR | Tools / fixes |
+|-----|----|---------------|
+| v0.2.2 | #21 | repo_create, repo_update, repo_mirror_push |
+| v0.2.3 | #22 | repo_tree, repo_topics_update, file_read dir fix |
+| v0.2.4 | #23 | issue_get, release_create, repo_delete |
+| v0.2.5 | #26 | repo_update archived+template, create_project_from_template template_name, pr_files_diff loop fix |
 
-#### #12 — repo_update: add `archived` and `template` fields
-**File:** `internal/gitea/repos.go` → `UpdateRepoArgs` struct
-**File:** `internal/tools/repo_update.go` → input schema + args struct
+Current main: `e31fd3f`. CI green. Deployed via Flux.
 
-Add to `UpdateRepoArgs`:
-```go
-Archived *bool
-Template *bool
-```
+### Next up
 
-Add to tool input schema:
-```json
-"archived": {
-  "type": "boolean",
-  "description": "Mark repo as archived (read-only). Requires confirm=<repo name>."
-},
-"template": {
-  "type": "boolean",
-  "description": "Toggle template repo flag."
-}
-```
+1. **`hyperguild new-project` v1** — primary next target.
+   See brain node `adr-new-project-gitea-first-github-mirror` for full flow spec.
 
-Add confirm-guard for `archived=true` (same pattern as `private=false`):
-```go
-if args.Archived != nil && *args.Archived {
-    if args.Confirm != args.Name {
-        return nil, fmt.Errorf("setting archived=true is irreversible: set confirm=%q to proceed", args.Name)
-    }
-}
-```
-
-New test cases to add in `repo_update_test.go`:
-- `TestRepoUpdateTool_Archive` — happy path with confirm
-- `TestRepoUpdateTool_ArchiveRequiresConfirm` — missing confirm returns error
-- `TestRepoUpdateTool_SetTemplate` — no confirm needed
-
-#### #24 — create_project_from_template: make template selectable
-**File:** `internal/tools/create_project_from_template.go`
-
-Add optional `template_name` param to input schema:
-```json
-"template_name": {
-  "type": "string",
-  "enum": ["template-go-web", "template-go-agent"],
-  "description": "Template repo to generate from. Defaults to template-go-web.",
-  "default": "template-go-web"
-}
-```
-
-The tool should use `args.TemplateName` if set, fall back to the hardcoded default.
-Remove the hardcoded template name from `cmd/gitea-mcp/main.go` constructor call —
-the tool resolves it internally.
-
-New test case: `TestCreateProjectFromTemplate_AgentTemplate`
-
-#### #25 — pr_files_diff: fix same diff returned for all files
-**File:** `internal/tools/pr_files_diff.go`
-
-There is a loop bug where all file entries in the response contain the same diff
-(the first file's diff is reused for every subsequent file). Find the loop and
-ensure each iteration reads and assigns the correct diff for its own file.
-
-Reproduce: call `pr_files_diff` on any PR with 3+ files, verify each file has
-a distinct diff.
-
-### Definition of done
-
-- [ ] `task check` passes
-- [ ] `repo_update` accepts `archived` and `template` params
-- [ ] `archived=true` requires `confirm=<repo name>`
-- [ ] `create_project_from_template` accepts `template_name` param, defaults to `template-go-web`
-- [ ] `pr_files_diff` returns distinct diff per file
-- [ ] All new test cases pass
-- [ ] PR `fix/v02-patch` merged to main via PR (not direct push)
-
-### After this sprint
-
-Next: `hyperguild new-project` v1 implementation.
-See brain node `adr-new-project-gitea-first-github-mirror` for the full flow spec.
-Also: verify end-to-end mirror flow (issue #19) once `repo_mirror_push` is confirmed working.
+2. **Issue #19** — end-to-end mirror flow verification.
+   `repo_mirror_push` is implemented but the full flow (create repo → add push mirror → verify sync to GitHub) has not been tested manually. Do this before relying on it in production.
 
 ---

.cursorrules

@@ -39,6 +39,18 @@ These rules apply to every task across every project, regardless of harness.
 4. **Goal-driven execution.** Define clear success criteria up front for every task.
    Loop — implement, verify, refine — until those criteria are met. Don't claim
    completion without evidence (tests pass, command output, observed behavior).
+5. **Trunk-Based Development — commit directly to main.** Every commit is one
+   logical change (one tool, one fix, one test) with passing tests. Main is always
+   deployable. Never create long-lived feature branches.
+
+   **Exception — parallel agents on same repo:** If another agent is known to be
+   actively working on the same repo simultaneously, create a short-lived branch
+   (`agent/<description>`), finish the task, and merge to main within the same
+   session. Do not leave agent branches open between sessions.
+
+   **Exception — external contributor or client four-eyes requirement:** Use
+   PR flow only when a human reviewer outside the project is required. Document
+   the reason in PROJECT.md.
 
 ## Default stack
 
@@ -52,6 +64,7 @@ These rules apply to every task across every project, regardless of harness.
 | Search | pgvector (vector), BM25 | Qdrant (when >1M vectors or hybrid retrieval) | — |
 | Logging | slog (structured) | — | — |
 | Testing | Table-driven, testify | — | — |
+| Agents (Go) | google.golang.org/adk + pkg/litellm adapter | — | — |
 
 Exploratory: Rust, Zig — I'll tell you when I want these.
 
@@ -61,9 +74,12 @@ Exploratory: Rust, Zig — I'll tell you when I want these.
 - **Errors**: `fmt.Errorf("operation: %w", err)` — never naked, never log-and-return
 - **Naming**: stdlib conventions, no stuttering
 - **Architecture**: prefer stdlib over frameworks, constructor injection, env-var config parsed into typed structs
-- **Git**: conventional commits (`feat:`, `fix:`, `chore:`), one concern per PR, PR describes *why* not *what*
+- **Git**: conventional commits (`feat:`, `fix:`, `chore:`), commit directly to main,
+  one logical change per commit, CI is the quality gate
+- **Never**: long-lived feature branches, PRs for solo work, direct push without
+  passing `task check` locally first
 - **Security**: no secrets in code, govulncheck before adding deps, SOPS for encrypted config
-- **Dependencies**: prefer stdlib. testify, slog, templ, sqlc are pre-approved; anything else needs justification in the commit message
+- **Dependencies**: prefer stdlib. testify, slog, templ, sqlc, google.golang.org/adk (agent projects only) are pre-approved; anything else needs justification in the commit message
 
 ## Infrastructure
 
@@ -103,18 +119,64 @@ See `~/dev/PROJECT_SUMMARY.md` for detailed descriptions of each project.
 - **koala-ai-stack** (`AGENTS/`) — local AI server infrastructure management
 - **klimatkollen** (`XT/`) — Swedish municipal climate data platform
 
-## Knowledge base
+## Knowledge base — actively use it
 
-When available, agents can query the shared knowledge base:
+A persistent brain (BM25 search + LLM-synthesised Q&A) survives across sessions,
+hosts, and harnesses. It holds 100+ hard-won entries: infra incident postmortems,
+Go pitfalls, framework gotchas, design principles, ADRs. **It is not optional
+reference material — query it actively, not just when explicitly told.**
 
-- **MCP**: `mcp://hyperguild.<TAILNET>.ts.net:3100/knowledge`
-- **HTTP**: `http://hyperguild.<TAILNET>.ts.net:3100/api/v1/search`
+### When to query (treat as a reflex)
 
-<!-- TODO: replace <TAILNET> placeholder with the real Tailscale tailnet
-     name once hyperguild is deployed. Until then, agents that try to
-     reach the knowledge service on a host where it isn't running will
-     get DNS NXDOMAIN, which is the desired fail-loudly behavior. -->
-- **Scoping**: defaults to `public` collection; client projects filter to `{client}` + `public`
+- **Before** starting a non-trivial task — search for prior art with the symptom
+  AND the system component ("how did we solve X in Y?"). 5 seconds beats 5 hours.
+- **When debugging** — search for the error string, the stack frame, the affected
+  service. Past you may have already paid this tax.
+- **Before adopting** a pattern, library, framework, or model name — check if it
+  was tried and rejected, or what the integration footguns are.
+- **When making architectural decisions** — search for the domain + "ADR" or
+  "decision" to find prior reasoning before re-deriving it.
+- **When a recommendation feels novel** — challenge yourself: "has this been
+  documented?" The brain often has it.
+
+### When to write
+
+After you discover something that **future-you would forget** and that **isn't
+recoverable from the code, git log, or PR description alone**:
+
+- Bugs whose root cause is non-obvious and generalisable beyond this project.
+- Framework / library / model-name quirks that bit you and would bite anyone.
+- Design principles validated under fire (e.g. "every `_get` needs a `_list`").
+- Postmortems for incidents: what broke, why, how diagnosed, what to do next time.
+
+DON'T write project status, sprint progress, PR summaries, or "what I did this
+session" — those rot fast and the originals are in git/gitea anyway. Brain
+entries that age well are about *why*, *how to avoid*, and *what to do when*.
+
+### How to access (per harness)
+
+| Harness | Query | Write |
+|---------|-------|-------|
+| **Claude Code, Claude Desktop** | `brain_query` (BM25), `brain_answer` (LLM-synth + sources) MCP tools | `brain_write` MCP tool |
+| **Crush, Pi, Antigravity, other MCP-capable** | same MCP server: `ingestion-brain` (via the `mcp__*_brain__*` namespace once authenticated) | same |
+| **Anything HTTP-only (curl, scripts)** | `POST https://brain-mcp.d-ma.be/query` with `{"query":"..."}` (auth via `BRAIN_MCP_TOKEN`) | `POST .../write` with `{"content":"...","filename":"..."}` |
+| **Browser / human inspection** | `https://gitea.d-ma.be/mathias/hyperguild` → `knowledge/` and `wiki/` markdown files |
+
+- **Scoping**: defaults to `public` collection; client projects filter to `{client}` + `public`.
+- **Routing**: brain_answer's LLM uses berget.ai as primary, iguana ollama as
+  fallback. Both are configurable in the `supervisor/ingestion-deployment.yaml`
+  on the koala k3s cluster; don't hardcode local-only model names into the
+  berget URL (see knowledge entry on namespace mismatches).
+
+### Quick reflex checks
+
+If you find yourself about to say any of these out loud, you owe yourself a brain query first:
+
+- "I think the issue might be..."
+- "Let me try X and see..."
+- "I'll just write a script to..."
+- "This is probably a new bug..."
+- "Has anyone done this before?" — *yes, probably, go check.*
 
 ## Client work rules
 
@@ -211,9 +273,11 @@ Key skills:
 
 ### Git
 - Conventional commits: `feat:`, `fix:`, `chore:`, `docs:`, `refactor:`
-- Branch naming: `feat/short-description`, `fix/short-description`
-- PRs: one concern per PR, description explains *why* not *what*
-- **Branch protection:** always work on a feature branch, open a PR, never push directly to main
+- **Trunk-Based Development:** commit directly to main. One logical change per commit.
+- Run `task check` locally before every push. CI is the quality gate, not branch protection.
+- No feature branches, no PRs for solo/agent work.
+- Exception: if a parallel agent session is active on this repo, use a short-lived
+  `agent/<description>` branch and merge within the same session.
 
 ### Security
 - No secrets in code, ever — use env vars or SOPS-encrypted files
@@ -251,98 +315,29 @@ When acting as a coding agent on this project:
 3. If unsure about a convention, check `DECISIONS.md` or ask
 4. Never modify files outside the project root without explicit permission
 5. When adding a dependency, explain why in the commit message
-6. Always work on a feature branch and open a PR — never push directly to main
+6. Commit directly to main. Run `task check` before every push. Never create
+   feature branches unless a parallel agent is simultaneously active on this repo.
 7. For client projects: never send code or context to cloud APIs — use local models via LiteLLM
 
-## Current sprint — gitea-mcp v0.2 patch (2026-05-14)
+## Current state — v0.2.5 (2026-05-17)
 
-### Context
-The main v0.2 batch (repo_create, repo_update, repo_mirror_push, repo_delete,
-repo_tree, repo_topics_update, file_read dir-fix, issue_get, release_create,
-create_project_from_template) was implemented and pushed directly to main.
+All v0.2 work is complete and deployed. No active sprint.
 
-This sprint fixes three remaining gaps found during code review on 2026-05-14.
-These are blockers for `hyperguild new-project`.
+### What shipped
 
-### Issues to fix (all three in one PR: `fix/v02-patch`)
+| Tag | PR | Tools / fixes |
+|-----|----|---------------|
+| v0.2.2 | #21 | repo_create, repo_update, repo_mirror_push |
+| v0.2.3 | #22 | repo_tree, repo_topics_update, file_read dir fix |
+| v0.2.4 | #23 | issue_get, release_create, repo_delete |
+| v0.2.5 | #26 | repo_update archived+template, create_project_from_template template_name, pr_files_diff loop fix |
 
-#### #12 — repo_update: add `archived` and `template` fields
-**File:** `internal/gitea/repos.go` → `UpdateRepoArgs` struct
-**File:** `internal/tools/repo_update.go` → input schema + args struct
+Current main: `e31fd3f`. CI green. Deployed via Flux.
 
-Add to `UpdateRepoArgs`:
-```go
-Archived *bool
-Template *bool
-```
+### Next up
 
-Add to tool input schema:
-```json
-"archived": {
-  "type": "boolean",
-  "description": "Mark repo as archived (read-only). Requires confirm=<repo name>."
-},
-"template": {
-  "type": "boolean",
-  "description": "Toggle template repo flag."
-}
-```
+1. **`hyperguild new-project` v1** — primary next target.
+   See brain node `adr-new-project-gitea-first-github-mirror` for full flow spec.
 
-Add confirm-guard for `archived=true` (same pattern as `private=false`):
-```go
-if args.Archived != nil && *args.Archived {
-    if args.Confirm != args.Name {
-        return nil, fmt.Errorf("setting archived=true is irreversible: set confirm=%q to proceed", args.Name)
-    }
-}
-```
-
-New test cases to add in `repo_update_test.go`:
-- `TestRepoUpdateTool_Archive` — happy path with confirm
-- `TestRepoUpdateTool_ArchiveRequiresConfirm` — missing confirm returns error
-- `TestRepoUpdateTool_SetTemplate` — no confirm needed
-
-#### #24 — create_project_from_template: make template selectable
-**File:** `internal/tools/create_project_from_template.go`
-
-Add optional `template_name` param to input schema:
-```json
-"template_name": {
-  "type": "string",
-  "enum": ["template-go-web", "template-go-agent"],
-  "description": "Template repo to generate from. Defaults to template-go-web.",
-  "default": "template-go-web"
-}
-```
-
-The tool should use `args.TemplateName` if set, fall back to the hardcoded default.
-Remove the hardcoded template name from `cmd/gitea-mcp/main.go` constructor call —
-the tool resolves it internally.
-
-New test case: `TestCreateProjectFromTemplate_AgentTemplate`
-
-#### #25 — pr_files_diff: fix same diff returned for all files
-**File:** `internal/tools/pr_files_diff.go`
-
-There is a loop bug where all file entries in the response contain the same diff
-(the first file's diff is reused for every subsequent file). Find the loop and
-ensure each iteration reads and assigns the correct diff for its own file.
-
-Reproduce: call `pr_files_diff` on any PR with 3+ files, verify each file has
-a distinct diff.
-
-### Definition of done
-
-- [ ] `task check` passes
-- [ ] `repo_update` accepts `archived` and `template` params
-- [ ] `archived=true` requires `confirm=<repo name>`
-- [ ] `create_project_from_template` accepts `template_name` param, defaults to `template-go-web`
-- [ ] `pr_files_diff` returns distinct diff per file
-- [ ] All new test cases pass
-- [ ] PR `fix/v02-patch` merged to main via PR (not direct push)
-
-### After this sprint
-
-Next: `hyperguild new-project` v1 implementation.
-See brain node `adr-new-project-gitea-first-github-mirror` for the full flow spec.
-Also: verify end-to-end mirror flow (issue #19) once `repo_mirror_push` is confirmed working.
+2. **Issue #19** — end-to-end mirror flow verification.
+   `repo_mirror_push` is implemented but the full flow (create repo → add push mirror → verify sync to GitHub) has not been tested manually. Do this before relying on it in production.

.gitea/workflows/cd.yml

@@ -1,11 +1,9 @@
 name: CD
 
-on:
+"on":
   push:
     branches: [main]
     tags: ["v*"]
-  pull_request:
-    branches: [main]
 
 env:
   IMAGE: gitea-mcp
@@ -43,7 +41,6 @@ jobs:
     name: Build & Import
     needs: check
     runs-on: self-hosted
-    if: github.event_name != 'pull_request'
     outputs:
       image-tag: ${{ steps.meta.outputs.sha-tag }}
     steps:

.githooks/pre-push

@@ -0,0 +1,5 @@
+#!/usr/bin/env bash
+set -euo pipefail
+echo "→ Running task check before push..."
+task check
+echo "✓ pre-push check passed"

AGENTS.md

@@ -36,6 +36,18 @@ These rules apply to every task across every project, regardless of harness.
 4. **Goal-driven execution.** Define clear success criteria up front for every task.
    Loop — implement, verify, refine — until those criteria are met. Don't claim
    completion without evidence (tests pass, command output, observed behavior).
+5. **Trunk-Based Development — commit directly to main.** Every commit is one
+   logical change (one tool, one fix, one test) with passing tests. Main is always
+   deployable. Never create long-lived feature branches.
+
+   **Exception — parallel agents on same repo:** If another agent is known to be
+   actively working on the same repo simultaneously, create a short-lived branch
+   (`agent/<description>`), finish the task, and merge to main within the same
+   session. Do not leave agent branches open between sessions.
+
+   **Exception — external contributor or client four-eyes requirement:** Use
+   PR flow only when a human reviewer outside the project is required. Document
+   the reason in PROJECT.md.
 
 ## Default stack
 
@@ -49,6 +61,7 @@ These rules apply to every task across every project, regardless of harness.
 | Search | pgvector (vector), BM25 | Qdrant (when >1M vectors or hybrid retrieval) | — |
 | Logging | slog (structured) | — | — |
 | Testing | Table-driven, testify | — | — |
+| Agents (Go) | google.golang.org/adk + pkg/litellm adapter | — | — |
 
 Exploratory: Rust, Zig — I'll tell you when I want these.
 
@@ -58,9 +71,12 @@ Exploratory: Rust, Zig — I'll tell you when I want these.
 - **Errors**: `fmt.Errorf("operation: %w", err)` — never naked, never log-and-return
 - **Naming**: stdlib conventions, no stuttering
 - **Architecture**: prefer stdlib over frameworks, constructor injection, env-var config parsed into typed structs
-- **Git**: conventional commits (`feat:`, `fix:`, `chore:`), one concern per PR, PR describes *why* not *what*
+- **Git**: conventional commits (`feat:`, `fix:`, `chore:`), commit directly to main,
+  one logical change per commit, CI is the quality gate
+- **Never**: long-lived feature branches, PRs for solo work, direct push without
+  passing `task check` locally first
 - **Security**: no secrets in code, govulncheck before adding deps, SOPS for encrypted config
-- **Dependencies**: prefer stdlib. testify, slog, templ, sqlc are pre-approved; anything else needs justification in the commit message
+- **Dependencies**: prefer stdlib. testify, slog, templ, sqlc, google.golang.org/adk (agent projects only) are pre-approved; anything else needs justification in the commit message
 
 ## Infrastructure
 
@@ -100,18 +116,64 @@ See `~/dev/PROJECT_SUMMARY.md` for detailed descriptions of each project.
 - **koala-ai-stack** (`AGENTS/`) — local AI server infrastructure management
 - **klimatkollen** (`XT/`) — Swedish municipal climate data platform
 
-## Knowledge base
+## Knowledge base — actively use it
 
-When available, agents can query the shared knowledge base:
+A persistent brain (BM25 search + LLM-synthesised Q&A) survives across sessions,
+hosts, and harnesses. It holds 100+ hard-won entries: infra incident postmortems,
+Go pitfalls, framework gotchas, design principles, ADRs. **It is not optional
+reference material — query it actively, not just when explicitly told.**
 
-- **MCP**: `mcp://hyperguild.<TAILNET>.ts.net:3100/knowledge`
-- **HTTP**: `http://hyperguild.<TAILNET>.ts.net:3100/api/v1/search`
+### When to query (treat as a reflex)
 
-<!-- TODO: replace <TAILNET> placeholder with the real Tailscale tailnet
-     name once hyperguild is deployed. Until then, agents that try to
-     reach the knowledge service on a host where it isn't running will
-     get DNS NXDOMAIN, which is the desired fail-loudly behavior. -->
-- **Scoping**: defaults to `public` collection; client projects filter to `{client}` + `public`
+- **Before** starting a non-trivial task — search for prior art with the symptom
+  AND the system component ("how did we solve X in Y?"). 5 seconds beats 5 hours.
+- **When debugging** — search for the error string, the stack frame, the affected
+  service. Past you may have already paid this tax.
+- **Before adopting** a pattern, library, framework, or model name — check if it
+  was tried and rejected, or what the integration footguns are.
+- **When making architectural decisions** — search for the domain + "ADR" or
+  "decision" to find prior reasoning before re-deriving it.
+- **When a recommendation feels novel** — challenge yourself: "has this been
+  documented?" The brain often has it.
+
+### When to write
+
+After you discover something that **future-you would forget** and that **isn't
+recoverable from the code, git log, or PR description alone**:
+
+- Bugs whose root cause is non-obvious and generalisable beyond this project.
+- Framework / library / model-name quirks that bit you and would bite anyone.
+- Design principles validated under fire (e.g. "every `_get` needs a `_list`").
+- Postmortems for incidents: what broke, why, how diagnosed, what to do next time.
+
+DON'T write project status, sprint progress, PR summaries, or "what I did this
+session" — those rot fast and the originals are in git/gitea anyway. Brain
+entries that age well are about *why*, *how to avoid*, and *what to do when*.
+
+### How to access (per harness)
+
+| Harness | Query | Write |
+|---------|-------|-------|
+| **Claude Code, Claude Desktop** | `brain_query` (BM25), `brain_answer` (LLM-synth + sources) MCP tools | `brain_write` MCP tool |
+| **Crush, Pi, Antigravity, other MCP-capable** | same MCP server: `ingestion-brain` (via the `mcp__*_brain__*` namespace once authenticated) | same |
+| **Anything HTTP-only (curl, scripts)** | `POST https://brain-mcp.d-ma.be/query` with `{"query":"..."}` (auth via `BRAIN_MCP_TOKEN`) | `POST .../write` with `{"content":"...","filename":"..."}` |
+| **Browser / human inspection** | `https://gitea.d-ma.be/mathias/hyperguild` → `knowledge/` and `wiki/` markdown files |
+
+- **Scoping**: defaults to `public` collection; client projects filter to `{client}` + `public`.
+- **Routing**: brain_answer's LLM uses berget.ai as primary, iguana ollama as
+  fallback. Both are configurable in the `supervisor/ingestion-deployment.yaml`
+  on the koala k3s cluster; don't hardcode local-only model names into the
+  berget URL (see knowledge entry on namespace mismatches).
+
+### Quick reflex checks
+
+If you find yourself about to say any of these out loud, you owe yourself a brain query first:
+
+- "I think the issue might be..."
+- "Let me try X and see..."
+- "I'll just write a script to..."
+- "This is probably a new bug..."
+- "Has anyone done this before?" — *yes, probably, go check.*
 
 ## Client work rules
 
@@ -208,9 +270,11 @@ Key skills:
 
 ### Git
 - Conventional commits: `feat:`, `fix:`, `chore:`, `docs:`, `refactor:`
-- Branch naming: `feat/short-description`, `fix/short-description`
-- PRs: one concern per PR, description explains *why* not *what*
-- **Branch protection:** always work on a feature branch, open a PR, never push directly to main
+- **Trunk-Based Development:** commit directly to main. One logical change per commit.
+- Run `task check` locally before every push. CI is the quality gate, not branch protection.
+- No feature branches, no PRs for solo/agent work.
+- Exception: if a parallel agent session is active on this repo, use a short-lived
+  `agent/<description>` branch and merge within the same session.
 
 ### Security
 - No secrets in code, ever — use env vars or SOPS-encrypted files
@@ -248,98 +312,29 @@ When acting as a coding agent on this project:
 3. If unsure about a convention, check `DECISIONS.md` or ask
 4. Never modify files outside the project root without explicit permission
 5. When adding a dependency, explain why in the commit message
-6. Always work on a feature branch and open a PR — never push directly to main
+6. Commit directly to main. Run `task check` before every push. Never create
+   feature branches unless a parallel agent is simultaneously active on this repo.
 7. For client projects: never send code or context to cloud APIs — use local models via LiteLLM
 
-## Current sprint — gitea-mcp v0.2 patch (2026-05-14)
+## Current state — v0.2.5 (2026-05-17)
 
-### Context
-The main v0.2 batch (repo_create, repo_update, repo_mirror_push, repo_delete,
-repo_tree, repo_topics_update, file_read dir-fix, issue_get, release_create,
-create_project_from_template) was implemented and pushed directly to main.
+All v0.2 work is complete and deployed. No active sprint.
 
-This sprint fixes three remaining gaps found during code review on 2026-05-14.
-These are blockers for `hyperguild new-project`.
+### What shipped
 
-### Issues to fix (all three in one PR: `fix/v02-patch`)
+| Tag | PR | Tools / fixes |
+|-----|----|---------------|
+| v0.2.2 | #21 | repo_create, repo_update, repo_mirror_push |
+| v0.2.3 | #22 | repo_tree, repo_topics_update, file_read dir fix |
+| v0.2.4 | #23 | issue_get, release_create, repo_delete |
+| v0.2.5 | #26 | repo_update archived+template, create_project_from_template template_name, pr_files_diff loop fix |
 
-#### #12 — repo_update: add `archived` and `template` fields
-**File:** `internal/gitea/repos.go` → `UpdateRepoArgs` struct
-**File:** `internal/tools/repo_update.go` → input schema + args struct
+Current main: `e31fd3f`. CI green. Deployed via Flux.
 
-Add to `UpdateRepoArgs`:
-```go
-Archived *bool
-Template *bool
-```
+### Next up
 
-Add to tool input schema:
-```json
-"archived": {
-  "type": "boolean",
-  "description": "Mark repo as archived (read-only). Requires confirm=<repo name>."
-},
-"template": {
-  "type": "boolean",
-  "description": "Toggle template repo flag."
-}
-```
+1. **`hyperguild new-project` v1** — primary next target.
+   See brain node `adr-new-project-gitea-first-github-mirror` for full flow spec.
 
-Add confirm-guard for `archived=true` (same pattern as `private=false`):
-```go
-if args.Archived != nil && *args.Archived {
-    if args.Confirm != args.Name {
-        return nil, fmt.Errorf("setting archived=true is irreversible: set confirm=%q to proceed", args.Name)
-    }
-}
-```
-
-New test cases to add in `repo_update_test.go`:
-- `TestRepoUpdateTool_Archive` — happy path with confirm
-- `TestRepoUpdateTool_ArchiveRequiresConfirm` — missing confirm returns error
-- `TestRepoUpdateTool_SetTemplate` — no confirm needed
-
-#### #24 — create_project_from_template: make template selectable
-**File:** `internal/tools/create_project_from_template.go`
-
-Add optional `template_name` param to input schema:
-```json
-"template_name": {
-  "type": "string",
-  "enum": ["template-go-web", "template-go-agent"],
-  "description": "Template repo to generate from. Defaults to template-go-web.",
-  "default": "template-go-web"
-}
-```
-
-The tool should use `args.TemplateName` if set, fall back to the hardcoded default.
-Remove the hardcoded template name from `cmd/gitea-mcp/main.go` constructor call —
-the tool resolves it internally.
-
-New test case: `TestCreateProjectFromTemplate_AgentTemplate`
-
-#### #25 — pr_files_diff: fix same diff returned for all files
-**File:** `internal/tools/pr_files_diff.go`
-
-There is a loop bug where all file entries in the response contain the same diff
-(the first file's diff is reused for every subsequent file). Find the loop and
-ensure each iteration reads and assigns the correct diff for its own file.
-
-Reproduce: call `pr_files_diff` on any PR with 3+ files, verify each file has
-a distinct diff.
-
-### Definition of done
-
-- [ ] `task check` passes
-- [ ] `repo_update` accepts `archived` and `template` params
-- [ ] `archived=true` requires `confirm=<repo name>`
-- [ ] `create_project_from_template` accepts `template_name` param, defaults to `template-go-web`
-- [ ] `pr_files_diff` returns distinct diff per file
-- [ ] All new test cases pass
-- [ ] PR `fix/v02-patch` merged to main via PR (not direct push)
-
-### After this sprint
-
-Next: `hyperguild new-project` v1 implementation.
-See brain node `adr-new-project-gitea-first-github-mirror` for the full flow spec.
-Also: verify end-to-end mirror flow (issue #19) once `repo_mirror_push` is confirmed working.
+2. **Issue #19** — end-to-end mirror flow verification.
+   `repo_mirror_push` is implemented but the full flow (create repo → add push mirror → verify sync to GitHub) has not been tested manually. Do this before relying on it in production.

CLAUDE.md

@@ -37,9 +37,11 @@
 
 ### Git
 - Conventional commits: `feat:`, `fix:`, `chore:`, `docs:`, `refactor:`
-- Branch naming: `feat/short-description`, `fix/short-description`
-- PRs: one concern per PR, description explains *why* not *what*
-- **Branch protection:** always work on a feature branch, open a PR, never push directly to main
+- **Trunk-Based Development:** commit directly to main. One logical change per commit.
+- Run `task check` locally before every push. CI is the quality gate, not branch protection.
+- No feature branches, no PRs for solo/agent work.
+- Exception: if a parallel agent session is active on this repo, use a short-lived
+  `agent/<description>` branch and merge within the same session.
 
 ### Security
 - No secrets in code, ever — use env vars or SOPS-encrypted files
@@ -77,98 +79,29 @@ When acting as a coding agent on this project:
 3. If unsure about a convention, check `DECISIONS.md` or ask
 4. Never modify files outside the project root without explicit permission
 5. When adding a dependency, explain why in the commit message
-6. Always work on a feature branch and open a PR — never push directly to main
+6. Commit directly to main. Run `task check` before every push. Never create
+   feature branches unless a parallel agent is simultaneously active on this repo.
 7. For client projects: never send code or context to cloud APIs — use local models via LiteLLM
 
-## Current sprint — gitea-mcp v0.2 patch (2026-05-14)
+## Current state — v0.2.5 (2026-05-17)
 
-### Context
-The main v0.2 batch (repo_create, repo_update, repo_mirror_push, repo_delete,
-repo_tree, repo_topics_update, file_read dir-fix, issue_get, release_create,
-create_project_from_template) was implemented and pushed directly to main.
+All v0.2 work is complete and deployed. No active sprint.
 
-This sprint fixes three remaining gaps found during code review on 2026-05-14.
-These are blockers for `hyperguild new-project`.
+### What shipped
 
-### Issues to fix (all three in one PR: `fix/v02-patch`)
+| Tag | PR | Tools / fixes |
+|-----|----|---------------|
+| v0.2.2 | #21 | repo_create, repo_update, repo_mirror_push |
+| v0.2.3 | #22 | repo_tree, repo_topics_update, file_read dir fix |
+| v0.2.4 | #23 | issue_get, release_create, repo_delete |
+| v0.2.5 | #26 | repo_update archived+template, create_project_from_template template_name, pr_files_diff loop fix |
 
-#### #12 — repo_update: add `archived` and `template` fields
-**File:** `internal/gitea/repos.go` → `UpdateRepoArgs` struct
-**File:** `internal/tools/repo_update.go` → input schema + args struct
+Current main: `e31fd3f`. CI green. Deployed via Flux.
 
-Add to `UpdateRepoArgs`:
-```go
-Archived *bool
-Template *bool
-```
+### Next up
 
-Add to tool input schema:
-```json
-"archived": {
-  "type": "boolean",
-  "description": "Mark repo as archived (read-only). Requires confirm=<repo name>."
-},
-"template": {
-  "type": "boolean",
-  "description": "Toggle template repo flag."
-}
-```
+1. **`hyperguild new-project` v1** — primary next target.
+   See brain node `adr-new-project-gitea-first-github-mirror` for full flow spec.
 
-Add confirm-guard for `archived=true` (same pattern as `private=false`):
-```go
-if args.Archived != nil && *args.Archived {
-    if args.Confirm != args.Name {
-        return nil, fmt.Errorf("setting archived=true is irreversible: set confirm=%q to proceed", args.Name)
-    }
-}
-```
-
-New test cases to add in `repo_update_test.go`:
-- `TestRepoUpdateTool_Archive` — happy path with confirm
-- `TestRepoUpdateTool_ArchiveRequiresConfirm` — missing confirm returns error
-- `TestRepoUpdateTool_SetTemplate` — no confirm needed
-
-#### #24 — create_project_from_template: make template selectable
-**File:** `internal/tools/create_project_from_template.go`
-
-Add optional `template_name` param to input schema:
-```json
-"template_name": {
-  "type": "string",
-  "enum": ["template-go-web", "template-go-agent"],
-  "description": "Template repo to generate from. Defaults to template-go-web.",
-  "default": "template-go-web"
-}
-```
-
-The tool should use `args.TemplateName` if set, fall back to the hardcoded default.
-Remove the hardcoded template name from `cmd/gitea-mcp/main.go` constructor call —
-the tool resolves it internally.
-
-New test case: `TestCreateProjectFromTemplate_AgentTemplate`
-
-#### #25 — pr_files_diff: fix same diff returned for all files
-**File:** `internal/tools/pr_files_diff.go`
-
-There is a loop bug where all file entries in the response contain the same diff
-(the first file's diff is reused for every subsequent file). Find the loop and
-ensure each iteration reads and assigns the correct diff for its own file.
-
-Reproduce: call `pr_files_diff` on any PR with 3+ files, verify each file has
-a distinct diff.
-
-### Definition of done
-
-- [ ] `task check` passes
-- [ ] `repo_update` accepts `archived` and `template` params
-- [ ] `archived=true` requires `confirm=<repo name>`
-- [ ] `create_project_from_template` accepts `template_name` param, defaults to `template-go-web`
-- [ ] `pr_files_diff` returns distinct diff per file
-- [ ] All new test cases pass
-- [ ] PR `fix/v02-patch` merged to main via PR (not direct push)
-
-### After this sprint
-
-Next: `hyperguild new-project` v1 implementation.
-See brain node `adr-new-project-gitea-first-github-mirror` for the full flow spec.
-Also: verify end-to-end mirror flow (issue #19) once `repo_mirror_push` is confirmed working.
+2. **Issue #19** — end-to-end mirror flow verification.
+   `repo_mirror_push` is implemented but the full flow (create repo → add push mirror → verify sync to GitHub) has not been tested manually. Do this before relying on it in production.

Dockerfile

@@ -1,5 +1,16 @@
 FROM golang:1.26-alpine AS build
 WORKDIR /src
+
+# Fetch internal gitea-hosted Go modules (e.g. mcp-chassis) without going
+# through proxy.golang.org and without HTTP→HTTPS surprises. Gitea returns
+# http:// in its go-import meta tag, so rewrite to https here and bypass
+# the module proxy + sumdb.
+RUN apk add --no-cache git && \
+    git config --global url."https://gitea.d-ma.be/".insteadOf "http://gitea.d-ma.be/"
+ENV GOPRIVATE=gitea.d-ma.be
+ENV GOPROXY=direct
+ENV GOSUMDB=off
+
 COPY go.mod go.sum ./
 RUN go mod download
 COPY . .

README.md

@@ -2,3 +2,14 @@
 
 Streamable HTTP MCP service exposing Gitea repo operations to Claude apps.
 See `~/dev/AI/infra/docs/superpowers/specs/2026-05-04-gitea-mcp-gitops-workflow-design.md`.
+
+## Quickstart
+
+```bash
+task setup:hooks   # installs .githooks/pre-push — runs task check before every push
+task check         # context sync + lint + test + vet
+task build         # produces bin/gitea-mcp
+```
+
+This repo uses Trunk-Based Development. Commit directly to `main`. The pre-push
+hook enforces the quality gate locally; CI re-runs `task check` on every push.

Taskfile.yml

@@ -47,6 +47,13 @@ tasks:
     cmds:
       - bash scripts/context-sync.sh
 
+  setup:hooks:
+    desc: Install git hooks (.githooks/pre-push)
+    cmds:
+      - git config core.hooksPath .githooks
+      - chmod +x .githooks/pre-push
+      - echo "✓ git hooks installed (pre-push runs task check)"
+
   context:sync:claude:
     cmds: [bash scripts/context-sync.sh claude]
   context:sync:agents:

cmd/gitea-mcp/healthz.go

@@ -0,0 +1,46 @@
+package main
+
+import (
+	"encoding/json"
+	"net/http"
+)
+
+type healthStatus struct {
+	OK  bool      `json:"ok"`
+	JWT jwtStatus `json:"jwt"`
+}
+
+// jwtStatus surfaces the runtime state of the Dex JWT validator so ops
+// can distinguish "Dex unreachable at startup" from "JWT auth not
+// configured" — both previously degraded silently to static-token-only
+// (refs hyperguild/gitea-mcp#6).
+type jwtStatus struct {
+	// Status is one of: "disabled" (DEX_ISSUER_URL not set),
+	// "enabled" (validator initialized), "degraded" (configured but
+	// init failed; only static-token auth currently accepted).
+	Status    string `json:"status"`
+	LastError string `json:"last_error,omitempty"`
+}
+
+func newHealthzHandler(dexConfigured, validatorReady bool, initErr error) http.HandlerFunc {
+	status := healthStatus{OK: true, JWT: jwtStatus{Status: jwtStatusFor(dexConfigured, validatorReady)}}
+	if initErr != nil {
+		status.JWT.LastError = initErr.Error()
+	}
+	body, _ := json.Marshal(status)
+	return func(w http.ResponseWriter, _ *http.Request) {
+		w.Header().Set("Content-Type", "application/json")
+		_, _ = w.Write(body)
+	}
+}
+
+func jwtStatusFor(dexConfigured, validatorReady bool) string {
+	switch {
+	case !dexConfigured:
+		return "disabled"
+	case validatorReady:
+		return "enabled"
+	default:
+		return "degraded"
+	}
+}

cmd/gitea-mcp/healthz_test.go

@@ -0,0 +1,60 @@
+package main
+
+import (
+	"encoding/json"
+	"errors"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func TestHealthzHandler(t *testing.T) {
+	tests := []struct {
+		name           string
+		dexConfigured  bool
+		validatorReady bool
+		initErr        error
+		wantStatus     string
+		wantLastError  string
+	}{
+		{
+			name:          "disabled when DEX_ISSUER_URL unset",
+			wantStatus:    "disabled",
+			wantLastError: "",
+		},
+		{
+			name:           "enabled when validator initialized",
+			dexConfigured:  true,
+			validatorReady: true,
+			wantStatus:     "enabled",
+			wantLastError:  "",
+		},
+		{
+			name:          "degraded when Dex configured but init failed",
+			dexConfigured: true,
+			initErr:       errors.New("fetch oidc discovery: dial tcp: connection refused"),
+			wantStatus:    "degraded",
+			wantLastError: "fetch oidc discovery: dial tcp: connection refused",
+		},
+	}
+
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			h := newHealthzHandler(tc.dexConfigured, tc.validatorReady, tc.initErr)
+			rec := httptest.NewRecorder()
+			h.ServeHTTP(rec, httptest.NewRequest(http.MethodGet, "/healthz", nil))
+
+			require.Equal(t, http.StatusOK, rec.Code)
+			assert.Equal(t, "application/json", rec.Header().Get("Content-Type"))
+
+			var got healthStatus
+			require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &got))
+			assert.True(t, got.OK)
+			assert.Equal(t, tc.wantStatus, got.JWT.Status)
+			assert.Equal(t, tc.wantLastError, got.JWT.LastError)
+		})
+	}
+}

cmd/gitea-mcp/main.go

@@ -2,10 +2,12 @@ package main
 
 import (
 	"context"
-	"encoding/json"
 	"log/slog"
 	"net/http"
 	"os"
+	"strings"
+
+	chassisauth "gitea.d-ma.be/mathias/mcp-chassis/auth"
 
 	"gitea.d-ma.be/mathias/gitea-mcp/internal/allowlist"
 	"gitea.d-ma.be/mathias/gitea-mcp/internal/auth"
@@ -27,78 +29,41 @@ func main() {
 
 	ctx := context.Background()
 
-	jwtValidator, err := auth.NewJWTValidator(ctx, cfg.DexIssuerURL, cfg.MCPAudience)
-	if err != nil {
-		logger.Warn("jwt validator init failed; JWT auth disabled", "err", err)
+	jwtValidator, jwtInitErr := chassisauth.NewJWTValidator(ctx, cfg.DexIssuerURL, cfg.MCPAudience)
+	if jwtInitErr != nil {
+		logger.Warn("jwt validator init failed; JWT auth degraded", "err", jwtInitErr)
 	}
 
 	giteaClient := gitea.NewClient(cfg.GiteaBaseURL, cfg.DefaultToken)
 	ownerAllow := allowlist.New(cfg.AllowedOwners)
 
 	reg := registry.New()
-	reg.Register(tools.NewRepoList(giteaClient, ownerAllow))
-	reg.Register(tools.NewRepoGet(giteaClient, ownerAllow))
-	reg.Register(tools.NewRepoSearch(giteaClient, ownerAllow))
-	reg.Register(tools.NewRepoStatus(giteaClient, ownerAllow))
-	reg.Register(tools.NewFileRead(giteaClient, ownerAllow))
-	reg.Register(tools.NewFileWriteBranch(giteaClient, ownerAllow))
-	reg.Register(tools.NewFileDelete(giteaClient, ownerAllow))
-	reg.Register(tools.NewDirList(giteaClient, ownerAllow))
-	reg.Register(tools.NewBranchList(giteaClient, ownerAllow))
-	reg.Register(tools.NewBranchDelete(giteaClient, ownerAllow))
-	reg.Register(tools.NewBranchProtectionGet(giteaClient, ownerAllow))
-	reg.Register(tools.NewPRCreate(giteaClient, ownerAllow))
-	reg.Register(tools.NewPRGet(giteaClient, ownerAllow))
-	reg.Register(tools.NewPRList(giteaClient, ownerAllow))
-	reg.Register(tools.NewPRMerge(giteaClient, ownerAllow))
-	reg.Register(tools.NewPRComment(giteaClient, ownerAllow))
-	reg.Register(tools.NewPRFilesDiff(giteaClient, ownerAllow))
-	reg.Register(tools.NewWorkflowRunTrigger(giteaClient, ownerAllow, cfg.GiteaBaseURL))
-	reg.Register(tools.NewWorkflowRunStatus(giteaClient, ownerAllow))
-	reg.Register(tools.NewCodeSearch(giteaClient, ownerAllow))
-	reg.Register(tools.NewIssueCreate(giteaClient, ownerAllow))
-	reg.Register(tools.NewIssueComment(giteaClient, ownerAllow))
-	reg.Register(tools.NewCreateProjectFromTemplate(giteaClient, ownerAllow, "mathias", "template-go-web"))
-	reg.Register(tools.NewTagCreate(giteaClient, ownerAllow))
-	reg.Register(tools.NewRepoCreate(giteaClient, ownerAllow))
-	reg.Register(tools.NewRepoUpdate(giteaClient, ownerAllow))
-	reg.Register(tools.NewRepoMirrorPush(giteaClient, ownerAllow))
-	reg.Register(tools.NewRepoTree(giteaClient, ownerAllow))
-	reg.Register(tools.NewRepoTopicsUpdate(giteaClient, ownerAllow))
-	reg.Register(tools.NewIssueGet(giteaClient, ownerAllow))
-	reg.Register(tools.NewReleaseCreate(giteaClient, ownerAllow))
-	reg.Register(tools.NewRepoDelete(giteaClient, ownerAllow))
+	tools.RegisterAll(reg, giteaClient, ownerAllow, cfg.GiteaBaseURL, "mathias", "template-go-web")
 
 	mcpSrv := mcp.NewServer(mcp.ServerOptions{
 		Registry: reg,
 		Sessions: mcp.NewSessionStore(),
 	})
 
+	// resourceMetadataURL is only emitted in the WWW-Authenticate challenge
+	// when both MCPResourceURL and a Dex issuer are wired; empty disables
+	// the challenge so static-only clients aren't pushed into OAuth discovery.
+	var resourceMetadataURL string
+	if cfg.MCPResourceURL != "" && cfg.DexIssuerURL != "" {
+		resourceMetadataURL = strings.TrimRight(cfg.MCPResourceURL, "/") + "/.well-known/oauth-protected-resource"
+	}
+
 	mux := http.NewServeMux()
 	mux.Handle("/mcp", mcp.OriginAllowlist(cfg.OriginAllowlist)(
-		auth.BearerMiddleware(jwtValidator, cfg.StaticToken,
+		chassisauth.BearerMiddleware(cfg.StaticToken, jwtValidator, "gitea", resourceMetadataURL,
 			auth.CallerMiddleware(mcpSrv),
 		),
 	))
-	mux.HandleFunc("/healthz", func(w http.ResponseWriter, _ *http.Request) {
-		w.WriteHeader(http.StatusOK)
-		_, _ = w.Write([]byte("ok"))
-	})
-	mux.HandleFunc("/.well-known/oauth-protected-resource", func(w http.ResponseWriter, r *http.Request) {
-		if r.Method != http.MethodGet {
-			http.Error(w, "method not allowed", http.StatusMethodNotAllowed)
-			return
-		}
-		w.Header().Set("Content-Type", "application/json")
-		payload := map[string]any{
-			"resource":             cfg.MCPResourceURL,
-			"authorization_servers": []string{},
-		}
+	mux.Handle("/healthz", newHealthzHandler(cfg.DexIssuerURL != "", jwtValidator != nil, jwtInitErr))
 	if cfg.DexIssuerURL != "" {
-			payload["authorization_servers"] = []string{cfg.DexIssuerURL}
+		mux.HandleFunc("GET /.well-known/oauth-protected-resource",
+			chassisauth.ProtectedResourceHandler(cfg.MCPResourceURL, cfg.DexIssuerURL))
 	}
-		_ = json.NewEncoder(w).Encode(payload)
-	})
 
 	addr := ":" + cfg.Port
 	logger.Info("gitea-mcp starting", "addr", addr, "version", "0.1.0")

go.mod

@@ -9,6 +9,7 @@ require (
 )
 
 require (
+	gitea.d-ma.be/mathias/mcp-chassis v0.1.0 // indirect
 	github.com/davecgh/go-spew v1.1.1 // indirect
 	github.com/decred/dcrd/dcrec/secp256k1/v4 v4.4.0 // indirect
 	github.com/goccy/go-json v0.10.3 // indirect

go.sum

@@ -1,3 +1,5 @@
+gitea.d-ma.be/mathias/mcp-chassis v0.1.0 h1:8RXO34+n7Vu8HnUMagars6fc4oemqRpMu7MVtjaj4qY=
+gitea.d-ma.be/mathias/mcp-chassis v0.1.0/go.mod h1:ajbLlwr2L7FAN3TBU39KucZkKJM02wTbKbDKDEW2YvE=
 github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=

internal/auth/bearer.go

@@ -1,42 +0,0 @@
-package auth
-
-import (
-	"crypto/subtle"
-	"net/http"
-	"strings"
-)
-
-// BearerMiddleware authenticates requests via the Authorization header.
-//
-// A request is allowed when:
-//
-//  1. The Bearer token is a valid JWT issued by the configured Dex OIDC server, or
-//  2. The Bearer token matches staticToken (constant-time compare).
-//
-// Any other case — including missing or empty Authorization header — returns 401.
-//
-// The Gitea service PAT is intentionally NOT used to authenticate the caller:
-// it is only used by the Gitea client for upstream API calls. Decoupling the
-// two prevents the MCP endpoint from being reachable anonymously when a service
-// PAT happens to be configured.
-func BearerMiddleware(jwtValidator *JWTValidator, staticToken string, next http.Handler) http.Handler {
-	return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-		bearer, hasBearer := strings.CutPrefix(r.Header.Get("Authorization"), "Bearer ")
-		if !hasBearer || bearer == "" {
-			http.Error(w, "unauthorized", http.StatusUnauthorized)
-			return
-		}
-
-		if jwtValidator.Validate(r.Context(), bearer) {
-			next.ServeHTTP(w, r)
-			return
-		}
-
-		if staticToken != "" && subtle.ConstantTimeCompare([]byte(bearer), []byte(staticToken)) == 1 {
-			next.ServeHTTP(w, r)
-			return
-		}
-
-		http.Error(w, "unauthorized", http.StatusUnauthorized)
-	})
-}

internal/auth/bearer_test.go

@@ -1,92 +0,0 @@
-package auth_test
-
-import (
-	"net/http"
-	"net/http/httptest"
-	"testing"
-
-	"gitea.d-ma.be/mathias/gitea-mcp/internal/auth"
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-)
-
-func okHandler(called *bool) http.Handler {
-	return http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
-		if called != nil {
-			*called = true
-		}
-		w.WriteHeader(http.StatusOK)
-	})
-}
-
-func TestBearerMiddleware_NoAuthHeader(t *testing.T) {
-	srv := httptest.NewServer(auth.BearerMiddleware(nil, "", okHandler(nil)))
-	defer srv.Close()
-
-	resp, err := http.Post(srv.URL+"/mcp", "application/json", nil)
-	require.NoError(t, err)
-	defer func() { _ = resp.Body.Close() }()
-	assert.Equal(t, http.StatusUnauthorized, resp.StatusCode)
-}
-
-func TestBearerMiddleware_NoAuthHeader_RejectsEvenWhenStaticConfigured(t *testing.T) {
-	// A configured staticToken must not allow unauthenticated callers through.
-	srv := httptest.NewServer(auth.BearerMiddleware(nil, "any-static", okHandler(nil)))
-	defer srv.Close()
-
-	resp, err := http.Post(srv.URL+"/mcp", "application/json", nil)
-	require.NoError(t, err)
-	defer func() { _ = resp.Body.Close() }()
-	assert.Equal(t, http.StatusUnauthorized, resp.StatusCode)
-}
-
-func TestBearerMiddleware_EmptyBearer(t *testing.T) {
-	srv := httptest.NewServer(auth.BearerMiddleware(nil, "static", okHandler(nil)))
-	defer srv.Close()
-
-	req, _ := http.NewRequest(http.MethodPost, srv.URL+"/mcp", nil)
-	req.Header.Set("Authorization", "Bearer ")
-	resp, err := http.DefaultClient.Do(req)
-	require.NoError(t, err)
-	defer func() { _ = resp.Body.Close() }()
-	assert.Equal(t, http.StatusUnauthorized, resp.StatusCode)
-}
-
-func TestBearerMiddleware_StaticToken_Valid(t *testing.T) {
-	const staticToken = "my-static-token"
-	called := false
-	srv := httptest.NewServer(auth.BearerMiddleware(nil, staticToken, okHandler(&called)))
-	defer srv.Close()
-
-	req, _ := http.NewRequest(http.MethodPost, srv.URL+"/mcp", nil)
-	req.Header.Set("Authorization", "Bearer "+staticToken)
-	resp, err := http.DefaultClient.Do(req)
-	require.NoError(t, err)
-	defer func() { _ = resp.Body.Close() }()
-	assert.Equal(t, http.StatusOK, resp.StatusCode)
-	assert.True(t, called)
-}
-
-func TestBearerMiddleware_StaticToken_Invalid(t *testing.T) {
-	srv := httptest.NewServer(auth.BearerMiddleware(nil, "correct-token", okHandler(nil)))
-	defer srv.Close()
-
-	req, _ := http.NewRequest(http.MethodPost, srv.URL+"/mcp", nil)
-	req.Header.Set("Authorization", "Bearer wrong-token")
-	resp, err := http.DefaultClient.Do(req)
-	require.NoError(t, err)
-	defer func() { _ = resp.Body.Close() }()
-	assert.Equal(t, http.StatusUnauthorized, resp.StatusCode)
-}
-
-func TestBearerMiddleware_UnknownBearer_NoStatic_NoJWT(t *testing.T) {
-	srv := httptest.NewServer(auth.BearerMiddleware(nil, "", okHandler(nil)))
-	defer srv.Close()
-
-	req, _ := http.NewRequest(http.MethodPost, srv.URL+"/mcp", nil)
-	req.Header.Set("Authorization", "Bearer random-unknown-token")
-	resp, err := http.DefaultClient.Do(req)
-	require.NoError(t, err)
-	defer func() { _ = resp.Body.Close() }()
-	assert.Equal(t, http.StatusUnauthorized, resp.StatusCode)
-}

internal/auth/jwt.go

@@ -1,79 +0,0 @@
-package auth
-
-import (
-	"context"
-	"encoding/json"
-	"fmt"
-	"net/http"
-	"time"
-
-	"github.com/lestrrat-go/jwx/v2/jwk"
-	"github.com/lestrrat-go/jwx/v2/jwt"
-)
-
-// JWTValidator validates bearer tokens as JWTs issued by a Dex OIDC server.
-// A nil JWTValidator always returns false — JWT validation is disabled.
-type JWTValidator struct {
-	issuer  string
-	aud     string
-	cache   *jwk.Cache
-	jwksURI string
-}
-
-// NewJWTValidator creates a validator by fetching the OIDC discovery document
-// from issuerURL. Returns nil, nil when issuerURL is empty (disabled).
-func NewJWTValidator(ctx context.Context, issuerURL, audience string) (*JWTValidator, error) {
-	if issuerURL == "" {
-		return nil, nil
-	}
-
-	resp, err := http.Get(issuerURL + "/.well-known/openid-configuration")
-	if err != nil {
-		return nil, fmt.Errorf("fetch oidc discovery: %w", err)
-	}
-	defer func() { _ = resp.Body.Close() }()
-
-	var doc struct {
-		JWKSURI string `json:"jwks_uri"`
-	}
-	if err := json.NewDecoder(resp.Body).Decode(&doc); err != nil {
-		return nil, fmt.Errorf("decode oidc discovery: %w", err)
-	}
-
-	cache := jwk.NewCache(ctx)
-	if err := cache.Register(doc.JWKSURI, jwk.WithRefreshInterval(time.Hour)); err != nil {
-		return nil, fmt.Errorf("register jwks uri: %w", err)
-	}
-	// warm the cache immediately so first request doesn't block
-	if _, err := cache.Refresh(ctx, doc.JWKSURI); err != nil {
-		return nil, fmt.Errorf("warm jwks cache: %w", err)
-	}
-
-	return &JWTValidator{
-		issuer:  issuerURL,
-		aud:     audience,
-		cache:   cache,
-		jwksURI: doc.JWKSURI,
-	}, nil
-}
-
-// Validate returns true if rawToken is a valid JWT signed by the OIDC server.
-func (v *JWTValidator) Validate(ctx context.Context, rawToken string) bool {
-	if v == nil {
-		return false
-	}
-	keySet, err := v.cache.Get(ctx, v.jwksURI)
-	if err != nil {
-		return false
-	}
-	opts := []jwt.ParseOption{
-		jwt.WithKeySet(keySet),
-		jwt.WithIssuer(v.issuer),
-		jwt.WithValidate(true),
-	}
-	if v.aud != "" {
-		opts = append(opts, jwt.WithAudience(v.aud))
-	}
-	_, err = jwt.Parse([]byte(rawToken), opts...)
-	return err == nil
-}

internal/gitea/client.go

@@ -3,8 +3,10 @@ package gitea
 import (
 	"bytes"
 	"context"
+	"fmt"
 	"io"
 	"net/http"
+	"strings"
 	"time"
 
 	"github.com/hashicorp/golang-lru/v2/expirable"
@@ -40,7 +42,21 @@ func (c *Client) DefaultBranch(ctx context.Context, owner, name string) (string,
 	return repo.DefaultBranch, nil
 }
 
+// hasEmptySegment reports whether the path portion (before any query string)
+// contains an empty segment ("//"), which means an owner or repo path
+// parameter was empty. Forwarding it upstream yields gitea's opaque
+// /api/swagger 404 (#36), so callers reject it locally instead.
+func hasEmptySegment(path string) bool {
+	if i := strings.IndexByte(path, '?'); i >= 0 {
+		path = path[:i]
+	}
+	return strings.Contains(path, "//")
+}
+
 func (c *Client) doOnce(ctx context.Context, method, path string, body []byte) ([]byte, int, error) {
+	if hasEmptySegment(path) {
+		return nil, 0, fmt.Errorf("%w: upstream path %q has an empty owner or repo segment", ErrValidation, path)
+	}
 	var reader io.Reader
 	if body != nil {
 		reader = bytes.NewReader(body)
@@ -107,6 +123,9 @@ type rawResponse struct {
 }
 
 func (c *Client) doRaw(ctx context.Context, method, path string, body []byte) (*rawResponse, error) {
+	if hasEmptySegment(path) {
+		return nil, fmt.Errorf("%w: upstream path %q has an empty owner or repo segment", ErrValidation, path)
+	}
 	var reader io.Reader
 	if body != nil {
 		reader = bytes.NewReader(body)

internal/gitea/client_guard_test.go

@@ -0,0 +1,54 @@
+package gitea_test
+
+import (
+	"context"
+	"net/http"
+	"net/http/httptest"
+	"sync/atomic"
+	"testing"
+
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/gitea"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+// #36 second line of defence: when owner or repo is empty the path contains an
+// empty segment ("//"). Rather than forward it upstream — where gitea answers
+// with its opaque /api/swagger 404 — the client must reject it locally with a
+// validation error and never touch the network.
+func TestEmptyPathSegmentRejectedBeforeNetwork(t *testing.T) {
+	var hits int32
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		atomic.AddInt32(&hits, 1)
+		w.WriteHeader(http.StatusOK)
+	}))
+	defer srv.Close()
+
+	c := gitea.NewClient(srv.URL, "tok")
+
+	paths := []string{
+		"/api/v1/repos/mathias//issues",        // empty repo
+		"/api/v1/repos//gitea-mcp/contents/",    // empty owner
+		"/api/v1/repos/mathias//issues?state=open", // empty repo before query
+	}
+	for _, p := range paths {
+		_, _, err := c.GetJSON(context.Background(), p)
+		require.Error(t, err, "path %q should be rejected", p)
+		assert.ErrorIs(t, err, gitea.ErrValidation)
+	}
+	assert.Equal(t, int32(0), atomic.LoadInt32(&hits), "guard must short-circuit before any HTTP call")
+}
+
+// A well-formed path with a query string must still pass the guard.
+func TestWellFormedPathPasses(t *testing.T) {
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		w.Header().Set("Content-Type", "application/json")
+		_, _ = w.Write([]byte(`[]`))
+	}))
+	defer srv.Close()
+
+	c := gitea.NewClient(srv.URL, "tok")
+	_, status, err := c.GetJSON(context.Background(), "/api/v1/repos/mathias/gitea-mcp/issues?state=open")
+	require.NoError(t, err)
+	assert.Equal(t, 200, status)
+}

internal/gitea/issues.go

@@ -4,6 +4,8 @@ import (
 	"context"
 	"encoding/json"
 	"fmt"
+	"net/url"
+	"strconv"
 )
 
 type Issue struct {
@@ -72,10 +74,128 @@ func (c *Client) CreateIssue(ctx context.Context, owner, repo string, args Creat
 	return &iss, nil
 }
 
+// ListIssuesArgs captures the optional query params for ListIssues.
+type ListIssuesArgs struct {
+	State  string // "open" | "closed" | "all"
+	Labels string // comma-separated label names
+	Since  string // ISO 8601
+	Page   int
+	Limit  int
+}
+
+// ListIssues fetches issues for a repo. Pulls are excluded server-side
+// (type=issues) so they don't leak through the same endpoint.
+func (c *Client) ListIssues(ctx context.Context, owner, repo string, args ListIssuesArgs) ([]Issue, error) {
+	q := url.Values{}
+	q.Set("type", "issues")
+	if args.State != "" {
+		q.Set("state", args.State)
+	}
+	if args.Labels != "" {
+		q.Set("labels", args.Labels)
+	}
+	if args.Since != "" {
+		q.Set("since", args.Since)
+	}
+	if args.Page > 0 {
+		q.Set("page", strconv.Itoa(args.Page))
+	}
+	if args.Limit > 0 {
+		q.Set("limit", strconv.Itoa(args.Limit))
+	}
+	p := fmt.Sprintf("/api/v1/repos/%s/%s/issues?%s", owner, repo, q.Encode())
+	body, status, err := c.GetJSON(ctx, p)
+	if err != nil {
+		return nil, err
+	}
+	if err := MapStatus(status, body); err != nil {
+		return nil, err
+	}
+	var issues []Issue
+	if err := json.Unmarshal(body, &issues); err != nil {
+		return nil, err
+	}
+	return issues, nil
+}
+
+// SetIssueState flips an issue between "open" and "closed" via PATCH.
+// Gitea uses the same endpoint for both transitions.
+func (c *Client) SetIssueState(ctx context.Context, owner, repo string, number int, state string) (*Issue, error) {
+	p := fmt.Sprintf("/api/v1/repos/%s/%s/issues/%d", owner, repo, number)
+	payload, err := json.Marshal(map[string]string{"state": state})
+	if err != nil {
+		return nil, err
+	}
+	body, status, err := c.PatchJSON(ctx, p, payload)
+	if err != nil {
+		return nil, err
+	}
+	if err := MapStatus(status, body); err != nil {
+		return nil, err
+	}
+	var iss Issue
+	if err := json.Unmarshal(body, &iss); err != nil {
+		return nil, err
+	}
+	return &iss, nil
+}
+
+// EditIssueArgs uses pointers so omitempty distinguishes "not set" (nil,
+// left untouched) from an explicit empty string (clears the field). Maps to
+// Gitea's PATCH /repos/{owner}/{repo}/issues/{index}.
+type EditIssueArgs struct {
+	Title *string `json:"title,omitempty"`
+	Body  *string `json:"body,omitempty"`
+}
+
+// EditIssue patches an issue's title and/or body. Only fields set in args are
+// sent, so omitted fields are left as-is server-side. Body is sent verbatim —
+// no identity footer — so repeated edits are idempotent.
+func (c *Client) EditIssue(ctx context.Context, owner, repo string, number int, args EditIssueArgs) (*Issue, error) {
+	p := fmt.Sprintf("/api/v1/repos/%s/%s/issues/%d", owner, repo, number)
+	payload, err := json.Marshal(args)
+	if err != nil {
+		return nil, err
+	}
+	body, status, err := c.PatchJSON(ctx, p, payload)
+	if err != nil {
+		return nil, err
+	}
+	if err := MapStatus(status, body); err != nil {
+		return nil, err
+	}
+	var iss Issue
+	if err := json.Unmarshal(body, &iss); err != nil {
+		return nil, err
+	}
+	return &iss, nil
+}
+
 type IssueComment struct {
 	ID        int64  `json:"id"`
 	Body      string `json:"body"`
 	HTMLURL   string `json:"html_url"`
+	User      User   `json:"user,omitempty"`
+	CreatedAt string `json:"created_at,omitempty"`
+	UpdatedAt string `json:"updated_at,omitempty"`
+}
+
+// ListIssueComments fetches all comments on an issue or pull request.
+// Per Gitea, /issues/{index}/comments serves both since PRs share index space with issues.
+func (c *Client) ListIssueComments(ctx context.Context, owner, repo string, index int) ([]IssueComment, error) {
+	p := fmt.Sprintf("/api/v1/repos/%s/%s/issues/%d/comments", owner, repo, index)
+	body, status, err := c.GetJSON(ctx, p)
+	if err != nil {
+		return nil, err
+	}
+	if err := MapStatus(status, body); err != nil {
+		return nil, err
+	}
+	var comments []IssueComment
+	if err := json.Unmarshal(body, &comments); err != nil {
+		return nil, err
+	}
+	return comments, nil
 }
 
 // CreateIssueComment posts to /issues/{index}/comments. Per Gitea, this same endpoint

internal/gitea/issues_test.go

@@ -76,6 +76,96 @@ func TestGetIssue_NotFound(t *testing.T) {
 	assert.ErrorIs(t, err, gitea.ErrNotFound)
 }
 
+func TestEditIssue(t *testing.T) {
+	var captured []byte
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		assert.Equal(t, http.MethodPatch, r.Method)
+		assert.Equal(t, "/api/v1/repos/o/r/issues/42", r.URL.Path)
+		var err error
+		captured, err = io.ReadAll(r.Body)
+		require.NoError(t, err)
+		w.Header().Set("Content-Type", "application/json")
+		_, _ = w.Write([]byte(`{"number":42,"title":"new title","body":"new body","state":"open","html_url":"http://example.com/issues/42"}`))
+	}))
+	defer srv.Close()
+
+	c := gitea.NewClient(srv.URL, "tok")
+	title, body := "new title", "new body"
+	iss, err := c.EditIssue(context.Background(), "o", "r", 42, gitea.EditIssueArgs{Title: &title, Body: &body})
+	require.NoError(t, err)
+
+	var payload map[string]any
+	require.NoError(t, json.Unmarshal(captured, &payload))
+	assert.Equal(t, "new title", payload["title"])
+	assert.Equal(t, "new body", payload["body"])
+
+	assert.Equal(t, 42, iss.Number)
+	assert.Equal(t, "new title", iss.Title)
+}
+
+// EditIssue must send only the fields explicitly provided — an omitted field
+// (nil pointer) is left untouched server-side.
+func TestEditIssue_PartialPatchOmitsUnsetFields(t *testing.T) {
+	var captured []byte
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		var err error
+		captured, err = io.ReadAll(r.Body)
+		require.NoError(t, err)
+		w.Header().Set("Content-Type", "application/json")
+		_, _ = w.Write([]byte(`{"number":42,"title":"only title","state":"open"}`))
+	}))
+	defer srv.Close()
+
+	c := gitea.NewClient(srv.URL, "tok")
+	title := "only title"
+	_, err := c.EditIssue(context.Background(), "o", "r", 42, gitea.EditIssueArgs{Title: &title})
+	require.NoError(t, err)
+
+	var payload map[string]any
+	require.NoError(t, json.Unmarshal(captured, &payload))
+	assert.Equal(t, "only title", payload["title"])
+	_, hasBody := payload["body"]
+	assert.False(t, hasBody, "body must be omitted when not set")
+}
+
+// An explicit empty-string body clears the field — pointer-to-"" is sent, not omitted.
+func TestEditIssue_EmptyBodyIsSent(t *testing.T) {
+	var captured []byte
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		var err error
+		captured, err = io.ReadAll(r.Body)
+		require.NoError(t, err)
+		w.Header().Set("Content-Type", "application/json")
+		_, _ = w.Write([]byte(`{"number":42,"body":"","state":"open"}`))
+	}))
+	defer srv.Close()
+
+	c := gitea.NewClient(srv.URL, "tok")
+	body := ""
+	_, err := c.EditIssue(context.Background(), "o", "r", 42, gitea.EditIssueArgs{Body: &body})
+	require.NoError(t, err)
+
+	var payload map[string]any
+	require.NoError(t, json.Unmarshal(captured, &payload))
+	val, hasBody := payload["body"]
+	assert.True(t, hasBody, "explicit empty body must be sent so it can clear the field")
+	assert.Equal(t, "", val)
+}
+
+func TestEditIssue_NotFound(t *testing.T) {
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		w.WriteHeader(http.StatusNotFound)
+		_, _ = w.Write([]byte(`{"message":"issue not found"}`))
+	}))
+	defer srv.Close()
+
+	c := gitea.NewClient(srv.URL, "tok")
+	title := "x"
+	_, err := c.EditIssue(context.Background(), "o", "r", 999, gitea.EditIssueArgs{Title: &title})
+	require.Error(t, err)
+	assert.ErrorIs(t, err, gitea.ErrNotFound)
+}
+
 func TestCreateIssueComment(t *testing.T) {
 	var captured []byte
 	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
@@ -101,3 +191,51 @@ func TestCreateIssueComment(t *testing.T) {
 	assert.Equal(t, "hello", comment.Body)
 	assert.Equal(t, "http://example.com/issues/42#comment-7", comment.HTMLURL)
 }
+
+func TestListIssueComments(t *testing.T) {
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		assert.Equal(t, http.MethodGet, r.Method)
+		assert.Equal(t, "/api/v1/repos/o/r/issues/42/comments", r.URL.Path)
+		w.Header().Set("Content-Type", "application/json")
+		_, _ = w.Write([]byte(`[
+			{"id":1,"body":"first","html_url":"http://example.com/issues/42#comment-1","user":{"login":"alice"},"created_at":"2026-05-01T00:00:00Z","updated_at":"2026-05-01T00:00:00Z"},
+			{"id":2,"body":"second","html_url":"http://example.com/issues/42#comment-2","user":{"login":"bob"},"created_at":"2026-05-02T00:00:00Z","updated_at":"2026-05-02T00:00:00Z"}
+		]`))
+	}))
+	defer srv.Close()
+
+	c := gitea.NewClient(srv.URL, "tok")
+	comments, err := c.ListIssueComments(context.Background(), "o", "r", 42)
+	require.NoError(t, err)
+	require.Len(t, comments, 2)
+	assert.Equal(t, int64(1), comments[0].ID)
+	assert.Equal(t, "first", comments[0].Body)
+	assert.Equal(t, "alice", comments[0].User.Login)
+	assert.Equal(t, "bob", comments[1].User.Login)
+}
+
+func TestListIssueComments_Empty(t *testing.T) {
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		w.Header().Set("Content-Type", "application/json")
+		_, _ = w.Write([]byte(`[]`))
+	}))
+	defer srv.Close()
+
+	c := gitea.NewClient(srv.URL, "tok")
+	comments, err := c.ListIssueComments(context.Background(), "o", "r", 42)
+	require.NoError(t, err)
+	assert.Empty(t, comments)
+}
+
+func TestListIssueComments_NotFound(t *testing.T) {
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		w.WriteHeader(http.StatusNotFound)
+		_, _ = w.Write([]byte(`{"message":"issue not found"}`))
+	}))
+	defer srv.Close()
+
+	c := gitea.NewClient(srv.URL, "tok")
+	_, err := c.ListIssueComments(context.Background(), "o", "r", 999)
+	require.Error(t, err)
+	assert.ErrorIs(t, err, gitea.ErrNotFound)
+}

internal/gitea/workflows.go

@@ -4,6 +4,7 @@ import (
 	"context"
 	"encoding/json"
 	"fmt"
+	"net/url"
 	"strconv"
 	"strings"
 )
@@ -55,10 +56,20 @@ func (c *Client) DispatchWorkflow(ctx context.Context, owner, repo, workflow str
 // WorkflowRun represents a Gitea Actions run.
 type WorkflowRun struct {
 	ID           int64  `json:"id"`
+	DisplayTitle string `json:"display_title,omitempty"`
 	Status       string `json:"status"`     // queued | in_progress | completed
 	Conclusion   string `json:"conclusion"` // success | failure | cancelled | skipped (only when completed)
+	Event        string `json:"event,omitempty"`
+	HeadSHA      string `json:"head_sha,omitempty"`
+	HeadBranch   string `json:"head_branch,omitempty"`
+	WorkflowID   string `json:"workflow_id,omitempty"`
+	RunNumber    int64  `json:"run_number,omitempty"`
 	StartedAt    string `json:"started_at"`
+	UpdatedAt    string `json:"updated_at,omitempty"`
 	HTMLURL      string `json:"html_url"`
+	Actor        struct {
+		Login string `json:"login"`
+	} `json:"actor,omitempty"`
 }
 
 // GetWorkflowRun fetches the status of a specific Actions run.
@@ -77,3 +88,59 @@ func (c *Client) GetWorkflowRun(ctx context.Context, owner, repo string, runID i
 	}
 	return &run, nil
 }
+
+// ListWorkflowRunsArgs captures the optional query params for ListWorkflowRuns.
+type ListWorkflowRunsArgs struct {
+	Branch   string
+	HeadSHA  string
+	Status   string // queued | in_progress | completed | all
+	Event    string // push | pull_request | schedule | workflow_dispatch | all
+	Workflow string
+	Page     int
+	Limit    int
+}
+
+type workflowRunsResponse struct {
+	TotalCount   int64         `json:"total_count"`
+	WorkflowRuns []WorkflowRun `json:"workflow_runs"`
+}
+
+// ListWorkflowRuns fetches recent Actions runs for a repo with optional filters.
+// Status / Event of "all" or "" are treated as no-filter.
+func (c *Client) ListWorkflowRuns(ctx context.Context, owner, repo string, args ListWorkflowRunsArgs) (*workflowRunsResponse, error) {
+	q := url.Values{}
+	if args.Branch != "" {
+		q.Set("branch", args.Branch)
+	}
+	if args.HeadSHA != "" {
+		q.Set("head_sha", args.HeadSHA)
+	}
+	if args.Status != "" && args.Status != "all" {
+		q.Set("status", args.Status)
+	}
+	if args.Event != "" && args.Event != "all" {
+		q.Set("event", args.Event)
+	}
+	if args.Workflow != "" {
+		q.Set("workflow", args.Workflow)
+	}
+	if args.Page > 0 {
+		q.Set("page", strconv.Itoa(args.Page))
+	}
+	if args.Limit > 0 {
+		q.Set("limit", strconv.Itoa(args.Limit))
+	}
+	p := fmt.Sprintf("/api/v1/repos/%s/%s/actions/runs?%s", owner, repo, q.Encode())
+	body, status, err := c.GetJSON(ctx, p)
+	if err != nil {
+		return nil, err
+	}
+	if err := MapStatus(status, body); err != nil {
+		return nil, err
+	}
+	var resp workflowRunsResponse
+	if err := json.Unmarshal(body, &resp); err != nil {
+		return nil, err
+	}
+	return &resp, nil
+}

internal/tools/alias_test.go

@@ -0,0 +1,49 @@
+package tools_test
+
+import (
+	"context"
+	"encoding/json"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/allowlist"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/gitea"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/tools"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+// #36: every caller (claude.ai connector, gitea's own convention) sends the
+// repo identifier as `repo` and issue/PR index as `index`, but the tools
+// declare them as `name` and `number`. Unmatched fields zero-valued the path
+// segment and produced gitea's misleading /api/swagger 404. parseArgs now
+// aliases repo->name and index->number so the idiomatic call works.
+func TestRepoAndIndexAliasesResolve(t *testing.T) {
+	tests := []struct {
+		name     string
+		args     string
+		wantPath string
+	}{
+		{"repo+index aliases", `{"owner":"mathias","repo":"infra","index":7}`, "/api/v1/repos/mathias/infra/issues/7"},
+		{"canonical name+number", `{"owner":"mathias","name":"infra","number":7}`, "/api/v1/repos/mathias/infra/issues/7"},
+		{"explicit name wins over repo", `{"owner":"mathias","name":"infra","repo":"ignored","number":7}`, "/api/v1/repos/mathias/infra/issues/7"},
+	}
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			var gotPath string
+			srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+				gotPath = r.URL.Path
+				w.Header().Set("Content-Type", "application/json")
+				_, _ = w.Write([]byte(`{"number":7,"title":"t"}`))
+			}))
+			defer srv.Close()
+
+			tool := tools.NewIssueGet(gitea.NewClient(srv.URL, "tok"), allowlist.New([]string{"mathias"}))
+			out, err := tool.Call(context.Background(), json.RawMessage(tc.args))
+			require.NoError(t, err)
+			assert.Equal(t, tc.wantPath, gotPath)
+			assert.Contains(t, string(out), `"number":7`)
+		})
+	}
+}

internal/tools/issue_close.go

@@ -0,0 +1,56 @@
+package tools
+
+import (
+	"context"
+	"encoding/json"
+
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/allowlist"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/gitea"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/registry"
+)
+
+type IssueClose struct {
+	c *gitea.Client
+	a *allowlist.Allowlist
+}
+
+func NewIssueClose(c *gitea.Client, a *allowlist.Allowlist) *IssueClose {
+	return &IssueClose{c: c, a: a}
+}
+
+func (t *IssueClose) Descriptor() registry.ToolDescriptor {
+	return registry.ToolDescriptor{
+		Name:        "issue_close",
+		Description: "Close an open issue. Reversible via issue_reopen.",
+		InputSchema: json.RawMessage(`{
+			"type":"object",
+			"properties":{
+				"owner":{"type":"string"},
+				"name":{"type":"string"},
+				"number":{"type":"integer","minimum":1}
+			},
+			"required":["owner","name","number"]
+		}`),
+	}
+}
+
+type issueCloseArgs struct {
+	Owner  string `json:"owner"`
+	Name   string `json:"name"`
+	Number int    `json:"number"`
+}
+
+func (t *IssueClose) Call(ctx context.Context, raw json.RawMessage) (json.RawMessage, error) {
+	var args issueCloseArgs
+	if err := parseArgs(raw, &args); err != nil {
+		return nil, err
+	}
+	if err := t.a.Check(args.Owner); err != nil {
+		return nil, err
+	}
+	iss, err := t.c.SetIssueState(ctx, args.Owner, args.Name, args.Number, "closed")
+	if err != nil {
+		return nil, err
+	}
+	return textOK(iss)
+}

internal/tools/issue_close_test.go

@@ -0,0 +1,52 @@
+package tools_test
+
+import (
+	"context"
+	"encoding/json"
+	"io"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/allowlist"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/gitea"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/tools"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func TestIssueCloseTool(t *testing.T) {
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		assert.Equal(t, http.MethodPatch, r.Method)
+		assert.Equal(t, "/api/v1/repos/mathias/infra/issues/26", r.URL.Path)
+		b, _ := io.ReadAll(r.Body)
+		assert.JSONEq(t, `{"state":"closed"}`, string(b))
+		w.Header().Set("Content-Type", "application/json")
+		_, _ = w.Write([]byte(`{"number":26,"title":"feat: ntfy via NPM","state":"closed","html_url":"http://gitea.example.com/mathias/infra/issues/26"}`))
+	}))
+	defer srv.Close()
+
+	tool := tools.NewIssueClose(gitea.NewClient(srv.URL, "tok"), allowlist.New([]string{"mathias"}))
+	out, err := tool.Call(context.Background(), json.RawMessage(`{"owner":"mathias","name":"infra","number":26}`))
+	require.NoError(t, err)
+	assert.Contains(t, string(out), `"number":26`)
+	assert.Contains(t, string(out), `"state":"closed"`)
+}
+
+func TestIssueCloseTool_NotFound(t *testing.T) {
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		w.WriteHeader(http.StatusNotFound)
+		_, _ = w.Write([]byte(`{"message":"issue not found"}`))
+	}))
+	defer srv.Close()
+
+	tool := tools.NewIssueClose(gitea.NewClient(srv.URL, "tok"), allowlist.New([]string{"mathias"}))
+	_, err := tool.Call(context.Background(), json.RawMessage(`{"owner":"mathias","name":"infra","number":999}`))
+	require.Error(t, err)
+}
+
+func TestIssueCloseAllowlistRejects(t *testing.T) {
+	tool := tools.NewIssueClose(gitea.NewClient("http://unused", ""), allowlist.New([]string{"mathias"}))
+	_, err := tool.Call(context.Background(), json.RawMessage(`{"owner":"evil","name":"x","number":1}`))
+	require.Error(t, err)
+}

internal/tools/issue_edit.go

@@ -0,0 +1,80 @@
+package tools
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/allowlist"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/gitea"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/registry"
+)
+
+type IssueEdit struct {
+	c *gitea.Client
+	a *allowlist.Allowlist
+}
+
+func NewIssueEdit(c *gitea.Client, a *allowlist.Allowlist) *IssueEdit {
+	return &IssueEdit{c: c, a: a}
+}
+
+func (t *IssueEdit) Descriptor() registry.ToolDescriptor {
+	return registry.ToolDescriptor{
+		Name: "issue_edit",
+		Description: "Edit an existing issue's title and/or body. Only fields explicitly set are patched; " +
+			"omitted fields are left untouched. Body is replaced verbatim (no identity footer) so edits are idempotent. " +
+			"WARNING: body is a full replacement — to amend rather than clobber, read-modify-write " +
+			"(fetch with issue_get, edit the text, send it back).",
+		InputSchema: json.RawMessage(`{
+			"type":"object",
+			"properties":{
+				"owner":{"type":"string"},
+				"name":{"type":"string"},
+				"number":{"type":"integer","minimum":1},
+				"title":{"type":"string","description":"New title. Omit to leave unchanged."},
+				"body":{"type":"string","description":"New body, full replacement. Omit to leave unchanged."}
+			},
+			"required":["owner","name","number"]
+		}`),
+	}
+}
+
+type issueEditArgs struct {
+	Owner  string  `json:"owner"`
+	Name   string  `json:"name"`
+	Number int     `json:"number"`
+	Title  *string `json:"title,omitempty"`
+	Body   *string `json:"body,omitempty"`
+}
+
+func (t *IssueEdit) Call(ctx context.Context, raw json.RawMessage) (json.RawMessage, error) {
+	var args issueEditArgs
+	if err := parseArgs(raw, &args); err != nil {
+		return nil, err
+	}
+	if err := t.a.Check(args.Owner); err != nil {
+		return nil, err
+	}
+	if args.Number < 1 {
+		return nil, fmt.Errorf("number is required: %w", gitea.ErrValidation)
+	}
+	if args.Title == nil && args.Body == nil {
+		return nil, fmt.Errorf("at least one of title or body must be set: %w", gitea.ErrValidation)
+	}
+
+	iss, err := t.c.EditIssue(ctx, args.Owner, args.Name, args.Number, gitea.EditIssueArgs{
+		Title: args.Title,
+		Body:  args.Body,
+	})
+	if err != nil {
+		return nil, err
+	}
+
+	return textOK(map[string]any{
+		"number":   iss.Number,
+		"title":    iss.Title,
+		"html_url": iss.HTMLURL,
+		"state":    iss.State,
+	})
+}

internal/tools/issue_edit_test.go

@@ -0,0 +1,91 @@
+package tools_test
+
+import (
+	"context"
+	"encoding/json"
+	"io"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/allowlist"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/gitea"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/tools"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func TestIssueEditTool(t *testing.T) {
+	var captured []byte
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		assert.Equal(t, http.MethodPatch, r.Method)
+		assert.Equal(t, "/api/v1/repos/mathias/infra/issues/26", r.URL.Path)
+		var err error
+		captured, err = io.ReadAll(r.Body)
+		require.NoError(t, err)
+		w.Header().Set("Content-Type", "application/json")
+		_, _ = w.Write([]byte(`{"number":26,"title":"new","state":"open","html_url":"http://gitea.example.com/mathias/infra/issues/26"}`))
+	}))
+	defer srv.Close()
+
+	tool := tools.NewIssueEdit(gitea.NewClient(srv.URL, "tok"), allowlist.New([]string{"mathias"}))
+	out, err := tool.Call(context.Background(), json.RawMessage(`{"owner":"mathias","name":"infra","number":26,"title":"new","body":"updated body"}`))
+	require.NoError(t, err)
+
+	var payload map[string]any
+	require.NoError(t, json.Unmarshal(captured, &payload))
+	assert.Equal(t, "new", payload["title"])
+	assert.Equal(t, "updated body", payload["body"])
+
+	assert.Contains(t, string(out), `"number":26`)
+}
+
+// Body must be sent verbatim — no identity footer appended (keeps edits idempotent).
+func TestIssueEditTool_BodyVerbatim(t *testing.T) {
+	var captured []byte
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		var err error
+		captured, err = io.ReadAll(r.Body)
+		require.NoError(t, err)
+		w.Header().Set("Content-Type", "application/json")
+		_, _ = w.Write([]byte(`{"number":26,"state":"open"}`))
+	}))
+	defer srv.Close()
+
+	tool := tools.NewIssueEdit(gitea.NewClient(srv.URL, "tok"), allowlist.New([]string{"mathias"}))
+	_, err := tool.Call(context.Background(), json.RawMessage(`{"owner":"mathias","name":"infra","number":26,"body":"exact text"}`))
+	require.NoError(t, err)
+
+	var payload map[string]any
+	require.NoError(t, json.Unmarshal(captured, &payload))
+	assert.Equal(t, "exact text", payload["body"], "body must be unchanged — no footer")
+	_, hasTitle := payload["title"]
+	assert.False(t, hasTitle, "title must be omitted when not provided")
+}
+
+func TestIssueEditTool_RequiresAField(t *testing.T) {
+	tool := tools.NewIssueEdit(gitea.NewClient("http://unused", ""), allowlist.New([]string{"mathias"}))
+	_, err := tool.Call(context.Background(), json.RawMessage(`{"owner":"mathias","name":"infra","number":26}`))
+	require.Error(t, err)
+	assert.ErrorIs(t, err, gitea.ErrValidation)
+}
+
+func TestIssueEditTool_NotFound(t *testing.T) {
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		w.WriteHeader(http.StatusNotFound)
+		_, _ = w.Write([]byte(`{"message":"issue not found"}`))
+	}))
+	defer srv.Close()
+
+	tool := tools.NewIssueEdit(gitea.NewClient(srv.URL, "tok"), allowlist.New([]string{"mathias"}))
+	title := "x"
+	body, _ := json.Marshal(map[string]any{"owner": "mathias", "name": "infra", "number": 999, "title": title})
+	_, err := tool.Call(context.Background(), body)
+	require.Error(t, err)
+}
+
+func TestIssueEditAllowlistRejects(t *testing.T) {
+	tool := tools.NewIssueEdit(gitea.NewClient("http://unused", ""), allowlist.New([]string{"mathias"}))
+	_, err := tool.Call(context.Background(), json.RawMessage(`{"owner":"evil","name":"x","number":1,"title":"y"}`))
+	require.Error(t, err)
+}

internal/tools/issue_list.go

@@ -0,0 +1,83 @@
+package tools
+
+import (
+	"context"
+	"encoding/json"
+
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/allowlist"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/gitea"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/registry"
+)
+
+type IssueList struct {
+	c *gitea.Client
+	a *allowlist.Allowlist
+}
+
+func NewIssueList(c *gitea.Client, a *allowlist.Allowlist) *IssueList {
+	return &IssueList{c: c, a: a}
+}
+
+func (t *IssueList) Descriptor() registry.ToolDescriptor {
+	return registry.ToolDescriptor{
+		Name:        "issue_list",
+		Description: "List issues in a repo with optional filters. PRs are excluded (use pr_list for those).",
+		InputSchema: json.RawMessage(`{
+			"type":"object",
+			"properties":{
+				"owner":{"type":"string"},
+				"name":{"type":"string"},
+				"state":{"type":"string","enum":["open","closed","all"]},
+				"labels":{"type":"string"},
+				"since":{"type":"string"},
+				"page":{"type":"integer","minimum":1},
+				"limit":{"type":"integer","minimum":1,"maximum":50}
+			},
+			"required":["owner","name"]
+		}`),
+	}
+}
+
+type issueListArgs struct {
+	Owner  string `json:"owner"`
+	Name   string `json:"name"`
+	State  string `json:"state"`
+	Labels string `json:"labels"`
+	Since  string `json:"since"`
+	Page   int    `json:"page"`
+	Limit  int    `json:"limit"`
+}
+
+func (t *IssueList) Call(ctx context.Context, raw json.RawMessage) (json.RawMessage, error) {
+	var args issueListArgs
+	if err := parseArgs(raw, &args); err != nil {
+		return nil, err
+	}
+	if err := t.a.Check(args.Owner); err != nil {
+		return nil, err
+	}
+	if args.State == "" {
+		args.State = "open"
+	}
+	args.Limit = capLimit(args.Limit, 30)
+	if args.Page < 1 {
+		args.Page = 1
+	}
+	issues, err := t.c.ListIssues(ctx, args.Owner, args.Name, gitea.ListIssuesArgs{
+		State:  args.State,
+		Labels: args.Labels,
+		Since:  args.Since,
+		Page:   args.Page,
+		Limit:  args.Limit,
+	})
+	if err != nil {
+		return nil, err
+	}
+	out := map[string]any{
+		"issues": issues,
+	}
+	if len(issues) == args.Limit {
+		out["next_page"] = args.Page + 1
+	}
+	return textOK(out)
+}

internal/tools/issue_list_comments.go

@@ -0,0 +1,56 @@
+package tools
+
+import (
+	"context"
+	"encoding/json"
+
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/allowlist"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/gitea"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/registry"
+)
+
+type IssueListComments struct {
+	c *gitea.Client
+	a *allowlist.Allowlist
+}
+
+func NewIssueListComments(c *gitea.Client, a *allowlist.Allowlist) *IssueListComments {
+	return &IssueListComments{c: c, a: a}
+}
+
+func (t *IssueListComments) Descriptor() registry.ToolDescriptor {
+	return registry.ToolDescriptor{
+		Name:        "issue_list_comments",
+		Description: "List all comments on an issue or pull request. Returns id, body, author, html_url, and timestamps for each comment.",
+		InputSchema: json.RawMessage(`{
+			"type":"object",
+			"properties":{
+				"owner":{"type":"string"},
+				"name":{"type":"string"},
+				"number":{"type":"integer","minimum":1}
+			},
+			"required":["owner","name","number"]
+		}`),
+	}
+}
+
+type issueListCommentsArgs struct {
+	Owner  string `json:"owner"`
+	Name   string `json:"name"`
+	Number int    `json:"number"`
+}
+
+func (t *IssueListComments) Call(ctx context.Context, raw json.RawMessage) (json.RawMessage, error) {
+	var args issueListCommentsArgs
+	if err := parseArgs(raw, &args); err != nil {
+		return nil, err
+	}
+	if err := t.a.Check(args.Owner); err != nil {
+		return nil, err
+	}
+	comments, err := t.c.ListIssueComments(ctx, args.Owner, args.Name, args.Number)
+	if err != nil {
+		return nil, err
+	}
+	return textOK(comments)
+}

internal/tools/issue_list_comments_test.go

@@ -0,0 +1,67 @@
+package tools_test
+
+import (
+	"context"
+	"encoding/json"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/allowlist"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/gitea"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/tools"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func TestIssueListCommentsTool(t *testing.T) {
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		assert.Equal(t, http.MethodGet, r.Method)
+		assert.Equal(t, "/api/v1/repos/mathias/infra/issues/42/comments", r.URL.Path)
+		w.Header().Set("Content-Type", "application/json")
+		_, _ = w.Write([]byte(`[
+			{"id":1,"body":"first","html_url":"http://gitea.example.com/mathias/infra/issues/42#comment-1","user":{"login":"alice"},"created_at":"2026-05-01T00:00:00Z","updated_at":"2026-05-01T00:00:00Z"},
+			{"id":2,"body":"second","html_url":"http://gitea.example.com/mathias/infra/issues/42#comment-2","user":{"login":"bob"},"created_at":"2026-05-02T00:00:00Z","updated_at":"2026-05-02T00:00:00Z"}
+		]`))
+	}))
+	defer srv.Close()
+
+	tool := tools.NewIssueListComments(gitea.NewClient(srv.URL, "tok"), allowlist.New([]string{"mathias"}))
+	out, err := tool.Call(context.Background(), json.RawMessage(`{"owner":"mathias","name":"infra","number":42}`))
+	require.NoError(t, err)
+	assert.Contains(t, string(out), `"id":1`)
+	assert.Contains(t, string(out), `"body":"first"`)
+	assert.Contains(t, string(out), `"login":"alice"`)
+	assert.Contains(t, string(out), `"login":"bob"`)
+}
+
+func TestIssueListCommentsTool_Empty(t *testing.T) {
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		w.Header().Set("Content-Type", "application/json")
+		_, _ = w.Write([]byte(`[]`))
+	}))
+	defer srv.Close()
+
+	tool := tools.NewIssueListComments(gitea.NewClient(srv.URL, "tok"), allowlist.New([]string{"mathias"}))
+	out, err := tool.Call(context.Background(), json.RawMessage(`{"owner":"mathias","name":"infra","number":42}`))
+	require.NoError(t, err)
+	assert.Contains(t, string(out), `[]`)
+}
+
+func TestIssueListCommentsTool_NotFound(t *testing.T) {
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		w.WriteHeader(http.StatusNotFound)
+		_, _ = w.Write([]byte(`{"message":"issue not found"}`))
+	}))
+	defer srv.Close()
+
+	tool := tools.NewIssueListComments(gitea.NewClient(srv.URL, "tok"), allowlist.New([]string{"mathias"}))
+	_, err := tool.Call(context.Background(), json.RawMessage(`{"owner":"mathias","name":"infra","number":999}`))
+	require.Error(t, err)
+}
+
+func TestIssueListCommentsAllowlistRejects(t *testing.T) {
+	tool := tools.NewIssueListComments(gitea.NewClient("http://unused", ""), allowlist.New([]string{"mathias"}))
+	_, err := tool.Call(context.Background(), json.RawMessage(`{"owner":"evil","name":"x","number":1}`))
+	require.Error(t, err)
+}

internal/tools/issue_list_test.go

@@ -0,0 +1,88 @@
+package tools_test
+
+import (
+	"context"
+	"encoding/json"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/allowlist"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/gitea"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/tools"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func TestIssueListTool(t *testing.T) {
+	tests := []struct {
+		name      string
+		input     string
+		wantQuery map[string]string
+		respBody  string
+		assert    func(t *testing.T, out string)
+	}{
+		{
+			name:      "happy path defaults",
+			input:     `{"owner":"mathias","name":"infra"}`,
+			wantQuery: map[string]string{"type": "issues", "state": "open", "page": "1", "limit": "30"},
+			respBody:  `[{"number":42,"title":"fix auth","state":"open","html_url":"http://gitea.example/m/infra/issues/42"},{"number":41,"title":"add tests","state":"open"}]`,
+			assert: func(t *testing.T, out string) {
+				assert.Contains(t, out, `"number":42`)
+				assert.Contains(t, out, `"number":41`)
+			},
+		},
+		{
+			name:      "state filter",
+			input:     `{"owner":"mathias","name":"infra","state":"closed"}`,
+			wantQuery: map[string]string{"type": "issues", "state": "closed"},
+			respBody:  `[]`,
+			assert: func(t *testing.T, out string) {
+				assert.Contains(t, out, `"issues":[]`)
+			},
+		},
+		{
+			name:      "label + since filter",
+			input:     `{"owner":"mathias","name":"infra","labels":"bug,critical","since":"2026-05-01T00:00:00Z"}`,
+			wantQuery: map[string]string{"labels": "bug,critical", "since": "2026-05-01T00:00:00Z"},
+			respBody:  `[]`,
+			assert:    func(t *testing.T, out string) {},
+		},
+		{
+			name:      "empty result",
+			input:     `{"owner":"mathias","name":"infra"}`,
+			wantQuery: map[string]string{"state": "open"},
+			respBody:  `[]`,
+			assert: func(t *testing.T, out string) {
+				assert.Contains(t, out, `"issues":[]`)
+				assert.NotContains(t, out, `next_page`)
+			},
+		},
+	}
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+				assert.Equal(t, http.MethodGet, r.Method)
+				assert.Equal(t, "/api/v1/repos/mathias/infra/issues", r.URL.Path)
+				q := r.URL.Query()
+				for k, v := range tc.wantQuery {
+					assert.Equal(t, v, q.Get(k), "query param %q", k)
+				}
+				w.Header().Set("Content-Type", "application/json")
+				_, _ = w.Write([]byte(tc.respBody))
+			}))
+			defer srv.Close()
+
+			tool := tools.NewIssueList(gitea.NewClient(srv.URL, "tok"), allowlist.New([]string{"mathias"}))
+			out, err := tool.Call(context.Background(), json.RawMessage(tc.input))
+			require.NoError(t, err)
+			tc.assert(t, string(out))
+		})
+	}
+}
+
+func TestIssueListAllowlistRejects(t *testing.T) {
+	tool := tools.NewIssueList(gitea.NewClient("http://unused", ""), allowlist.New([]string{"mathias"}))
+	_, err := tool.Call(context.Background(), json.RawMessage(`{"owner":"evil","name":"x"}`))
+	require.Error(t, err)
+}

internal/tools/issue_reopen.go

@@ -0,0 +1,56 @@
+package tools
+
+import (
+	"context"
+	"encoding/json"
+
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/allowlist"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/gitea"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/registry"
+)
+
+type IssueReopen struct {
+	c *gitea.Client
+	a *allowlist.Allowlist
+}
+
+func NewIssueReopen(c *gitea.Client, a *allowlist.Allowlist) *IssueReopen {
+	return &IssueReopen{c: c, a: a}
+}
+
+func (t *IssueReopen) Descriptor() registry.ToolDescriptor {
+	return registry.ToolDescriptor{
+		Name:        "issue_reopen",
+		Description: "Reopen a closed issue. Reversible via issue_close.",
+		InputSchema: json.RawMessage(`{
+			"type":"object",
+			"properties":{
+				"owner":{"type":"string"},
+				"name":{"type":"string"},
+				"number":{"type":"integer","minimum":1}
+			},
+			"required":["owner","name","number"]
+		}`),
+	}
+}
+
+type issueReopenArgs struct {
+	Owner  string `json:"owner"`
+	Name   string `json:"name"`
+	Number int    `json:"number"`
+}
+
+func (t *IssueReopen) Call(ctx context.Context, raw json.RawMessage) (json.RawMessage, error) {
+	var args issueReopenArgs
+	if err := parseArgs(raw, &args); err != nil {
+		return nil, err
+	}
+	if err := t.a.Check(args.Owner); err != nil {
+		return nil, err
+	}
+	iss, err := t.c.SetIssueState(ctx, args.Owner, args.Name, args.Number, "open")
+	if err != nil {
+		return nil, err
+	}
+	return textOK(iss)
+}

internal/tools/issue_reopen_test.go

@@ -0,0 +1,40 @@
+package tools_test
+
+import (
+	"context"
+	"encoding/json"
+	"io"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/allowlist"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/gitea"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/tools"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func TestIssueReopenTool(t *testing.T) {
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		assert.Equal(t, http.MethodPatch, r.Method)
+		assert.Equal(t, "/api/v1/repos/mathias/infra/issues/26", r.URL.Path)
+		b, _ := io.ReadAll(r.Body)
+		assert.JSONEq(t, `{"state":"open"}`, string(b))
+		w.Header().Set("Content-Type", "application/json")
+		_, _ = w.Write([]byte(`{"number":26,"title":"feat: ntfy via NPM","state":"open","html_url":"http://gitea.example.com/mathias/infra/issues/26"}`))
+	}))
+	defer srv.Close()
+
+	tool := tools.NewIssueReopen(gitea.NewClient(srv.URL, "tok"), allowlist.New([]string{"mathias"}))
+	out, err := tool.Call(context.Background(), json.RawMessage(`{"owner":"mathias","name":"infra","number":26}`))
+	require.NoError(t, err)
+	assert.Contains(t, string(out), `"number":26`)
+	assert.Contains(t, string(out), `"state":"open"`)
+}
+
+func TestIssueReopenAllowlistRejects(t *testing.T) {
+	tool := tools.NewIssueReopen(gitea.NewClient("http://unused", ""), allowlist.New([]string{"mathias"}))
+	_, err := tool.Call(context.Background(), json.RawMessage(`{"owner":"evil","name":"x","number":1}`))
+	require.Error(t, err)
+}

internal/tools/registry.go

@@ -0,0 +1,59 @@
+package tools
+
+import (
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/allowlist"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/gitea"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/registry"
+)
+
+// RegisterAll registers every gitea-mcp tool on reg. main.go and the
+// dispatch round-trip test share this single list so a newly added tool
+// cannot be wired in one place but missing from the other.
+//
+// giteaBaseURL is needed by workflow_run_trigger; tmplOwner/tmplRepo seed
+// create_project_from_template's default source template.
+func RegisterAll(
+	reg *registry.Registry,
+	c *gitea.Client,
+	a *allowlist.Allowlist,
+	giteaBaseURL, tmplOwner, tmplRepo string,
+) {
+	reg.Register(NewRepoList(c, a))
+	reg.Register(NewRepoGet(c, a))
+	reg.Register(NewRepoSearch(c, a))
+	reg.Register(NewRepoStatus(c, a))
+	reg.Register(NewFileRead(c, a))
+	reg.Register(NewFileWriteBranch(c, a))
+	reg.Register(NewFileDelete(c, a))
+	reg.Register(NewDirList(c, a))
+	reg.Register(NewBranchList(c, a))
+	reg.Register(NewBranchDelete(c, a))
+	reg.Register(NewBranchProtectionGet(c, a))
+	reg.Register(NewPRCreate(c, a))
+	reg.Register(NewPRGet(c, a))
+	reg.Register(NewPRList(c, a))
+	reg.Register(NewPRMerge(c, a))
+	reg.Register(NewPRComment(c, a))
+	reg.Register(NewPRFilesDiff(c, a))
+	reg.Register(NewWorkflowRunTrigger(c, a, giteaBaseURL))
+	reg.Register(NewWorkflowRunStatus(c, a))
+	reg.Register(NewCodeSearch(c, a))
+	reg.Register(NewIssueCreate(c, a))
+	reg.Register(NewIssueEdit(c, a))
+	reg.Register(NewIssueComment(c, a))
+	reg.Register(NewCreateProjectFromTemplate(c, a, tmplOwner, tmplRepo))
+	reg.Register(NewTagCreate(c, a))
+	reg.Register(NewRepoCreate(c, a))
+	reg.Register(NewRepoUpdate(c, a))
+	reg.Register(NewRepoMirrorPush(c, a))
+	reg.Register(NewRepoTree(c, a))
+	reg.Register(NewRepoTopicsUpdate(c, a))
+	reg.Register(NewIssueGet(c, a))
+	reg.Register(NewIssueList(c, a))
+	reg.Register(NewIssueListComments(c, a))
+	reg.Register(NewIssueClose(c, a))
+	reg.Register(NewIssueReopen(c, a))
+	reg.Register(NewWorkflowRunList(c, a))
+	reg.Register(NewReleaseCreate(c, a))
+	reg.Register(NewRepoDelete(c, a))
+}

internal/tools/registry_test.go

@@ -0,0 +1,58 @@
+package tools_test
+
+import (
+	"context"
+	"encoding/json"
+	"errors"
+	"testing"
+
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/allowlist"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/gitea"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/registry"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/tools"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+// buildRegistry wires the full tool set exactly as main.go does, against an
+// unroutable gitea base URL so any handler that reaches the network fails fast
+// (connection refused) rather than hanging.
+func buildRegistry() *registry.Registry {
+	reg := registry.New()
+	c := gitea.NewClient("http://127.0.0.1:1", "")
+	a := allowlist.New([]string{"mathias"})
+	tools.RegisterAll(reg, c, a, "http://127.0.0.1:1", "mathias", "template-go-web")
+	return reg
+}
+
+// Every registered tool must be dispatchable and advertise a parseable input
+// schema. This is the regression guard for #36's whole class: a tool that is
+// wired up but unroutable (or ships a malformed schema) fails CI here instead
+// of 404ing a live caller.
+func TestEveryRegisteredToolIsDispatchable(t *testing.T) {
+	reg := buildRegistry()
+	descs := reg.Tools()
+	require.NotEmpty(t, descs)
+
+	for _, d := range descs {
+		t.Run(d.Name, func(t *testing.T) {
+			require.NotEmpty(t, d.Name, "tool has empty name")
+			assert.True(t, json.Valid(d.InputSchema),
+				"tool %q ships invalid JSON input schema", d.Name)
+
+			// Dispatch with empty args. We do not care whether the call
+			// succeeds (most fail allowlist/validation/network) — only that
+			// the name resolves to a handler. ErrToolNotFound here means the
+			// tool advertised a name the dispatcher cannot route.
+			_, err := reg.Dispatch(context.Background(), d.Name, json.RawMessage(`{}`))
+			assert.False(t, errors.Is(err, registry.ErrToolNotFound),
+				"registered tool %q does not dispatch", d.Name)
+		})
+	}
+}
+
+// Lock the tool count so an accidental drop of a registration in RegisterAll
+// (the single source main.go and this test share) fails loudly.
+func TestRegisteredToolCount(t *testing.T) {
+	assert.Len(t, buildRegistry().Tools(), 38)
+}

internal/tools/tool.go

@@ -2,6 +2,7 @@ package tools
 
 import (
 	"encoding/json"
+	"strings"
 
 	"gitea.d-ma.be/mathias/gitea-mcp/internal/registry"
 )
@@ -17,7 +18,54 @@ func parseArgs(raw json.RawMessage, dst any) error {
 	if len(raw) == 0 {
 		return json.Unmarshal([]byte("{}"), dst)
 	}
-	return json.Unmarshal(raw, dst)
+	return json.Unmarshal(normalizeAliases(raw), dst)
+}
+
+// normalizeAliases maps the near-universal gitea/GitHub argument names onto the
+// idiosyncratic ones these tools declare: `repo` -> `name`, `index` -> `number`.
+// Every MCP caller (the claude.ai connector, LLMs primed on gitea's own API)
+// reaches for `repo`/`index`; without this the unmatched fields zero-valued the
+// upstream path segment and produced gitea's misleading /api/swagger 404 (#36).
+// The alias key is left intact so a tool whose real field IS `index`
+// (e.g. pr_merge) is unaffected.
+func normalizeAliases(raw json.RawMessage) json.RawMessage {
+	var m map[string]json.RawMessage
+	if err := json.Unmarshal(raw, &m); err != nil {
+		return raw // not a JSON object — leave untouched
+	}
+	changed := aliasInto(m, "name", "repo")
+	changed = aliasInto(m, "number", "index") || changed
+	if !changed {
+		return raw
+	}
+	if patched, err := json.Marshal(m); err == nil {
+		return patched
+	}
+	return raw
+}
+
+// aliasInto copies m[alias] into m[canonical] when canonical is absent or an
+// empty/zero JSON value and alias carries a usable one. An explicit canonical
+// value always wins. Returns true if it wrote canonical.
+func aliasInto(m map[string]json.RawMessage, canonical, alias string) bool {
+	av, ok := m[alias]
+	if !ok || isEmptyJSON(av) {
+		return false
+	}
+	if cv, ok := m[canonical]; ok && !isEmptyJSON(cv) {
+		return false
+	}
+	m[canonical] = av
+	return true
+}
+
+func isEmptyJSON(v json.RawMessage) bool {
+	switch strings.TrimSpace(string(v)) {
+	case "", `""`, "null", "0":
+		return true
+	default:
+		return false
+	}
 }
 
 // capLimit returns a sane page size: 0 or negative → def, > 50 → 50.

internal/tools/workflow_run_list.go

@@ -0,0 +1,87 @@
+package tools
+
+import (
+	"context"
+	"encoding/json"
+
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/allowlist"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/gitea"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/registry"
+)
+
+type WorkflowRunList struct {
+	c *gitea.Client
+	a *allowlist.Allowlist
+}
+
+func NewWorkflowRunList(c *gitea.Client, a *allowlist.Allowlist) *WorkflowRunList {
+	return &WorkflowRunList{c: c, a: a}
+}
+
+func (t *WorkflowRunList) Descriptor() registry.ToolDescriptor {
+	return registry.ToolDescriptor{
+		Name:        "workflow_run_list",
+		Description: "List recent Gitea Actions workflow runs with optional filters (branch, head_sha, status, event, workflow).",
+		InputSchema: json.RawMessage(`{
+			"type":"object",
+			"properties":{
+				"owner":{"type":"string"},
+				"name":{"type":"string"},
+				"branch":{"type":"string"},
+				"head_sha":{"type":"string"},
+				"status":{"type":"string","enum":["queued","in_progress","completed","all"]},
+				"event":{"type":"string","enum":["push","pull_request","schedule","workflow_dispatch","all"]},
+				"workflow":{"type":"string"},
+				"page":{"type":"integer","minimum":1},
+				"limit":{"type":"integer","minimum":1,"maximum":50}
+			},
+			"required":["owner","name"]
+		}`),
+	}
+}
+
+type workflowRunListArgs struct {
+	Owner    string `json:"owner"`
+	Name     string `json:"name"`
+	Branch   string `json:"branch"`
+	HeadSHA  string `json:"head_sha"`
+	Status   string `json:"status"`
+	Event    string `json:"event"`
+	Workflow string `json:"workflow"`
+	Page     int    `json:"page"`
+	Limit    int    `json:"limit"`
+}
+
+func (t *WorkflowRunList) Call(ctx context.Context, raw json.RawMessage) (json.RawMessage, error) {
+	var args workflowRunListArgs
+	if err := parseArgs(raw, &args); err != nil {
+		return nil, err
+	}
+	if err := t.a.Check(args.Owner); err != nil {
+		return nil, err
+	}
+	args.Limit = capLimit(args.Limit, 10)
+	if args.Page < 1 {
+		args.Page = 1
+	}
+	resp, err := t.c.ListWorkflowRuns(ctx, args.Owner, args.Name, gitea.ListWorkflowRunsArgs{
+		Branch:   args.Branch,
+		HeadSHA:  args.HeadSHA,
+		Status:   args.Status,
+		Event:    args.Event,
+		Workflow: args.Workflow,
+		Page:     args.Page,
+		Limit:    args.Limit,
+	})
+	if err != nil {
+		return nil, err
+	}
+	out := map[string]any{
+		"runs":  resp.WorkflowRuns,
+		"total": resp.TotalCount,
+	}
+	if len(resp.WorkflowRuns) == args.Limit {
+		out["next_page"] = args.Page + 1
+	}
+	return textOK(out)
+}

internal/tools/workflow_run_list_test.go

@@ -0,0 +1,98 @@
+package tools_test
+
+import (
+	"context"
+	"encoding/json"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/allowlist"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/gitea"
+	"gitea.d-ma.be/mathias/gitea-mcp/internal/tools"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func TestWorkflowRunListTool(t *testing.T) {
+	tests := []struct {
+		name      string
+		input     string
+		wantQuery map[string]string
+		notQuery  []string
+		respBody  string
+		assert    func(t *testing.T, out string)
+	}{
+		{
+			name:      "happy path defaults",
+			input:     `{"owner":"mathias","name":"gitea-mcp"}`,
+			wantQuery: map[string]string{"page": "1", "limit": "10"},
+			respBody:  `{"total_count":2,"workflow_runs":[{"id":823,"status":"completed","conclusion":"success","head_sha":"dc907fb"},{"id":822,"status":"completed","conclusion":"success","head_sha":"c4bd339"}]}`,
+			assert: func(t *testing.T, out string) {
+				assert.Contains(t, out, `"id":823`)
+				assert.Contains(t, out, `"total":2`)
+			},
+		},
+		{
+			name:      "head_sha short filter",
+			input:     `{"owner":"mathias","name":"gitea-mcp","head_sha":"dc907fb"}`,
+			wantQuery: map[string]string{"head_sha": "dc907fb"},
+			respBody:  `{"total_count":1,"workflow_runs":[{"id":823,"status":"completed","conclusion":"success","head_sha":"dc907fb"}]}`,
+			assert: func(t *testing.T, out string) {
+				assert.Contains(t, out, `"id":823`)
+			},
+		},
+		{
+			name:      "status filter",
+			input:     `{"owner":"mathias","name":"gitea-mcp","status":"in_progress"}`,
+			wantQuery: map[string]string{"status": "in_progress"},
+			respBody:  `{"total_count":0,"workflow_runs":[]}`,
+			assert: func(t *testing.T, out string) {
+				assert.Contains(t, out, `"runs":[]`)
+			},
+		},
+		{
+			name:      "status=all is no-op",
+			input:     `{"owner":"mathias","name":"gitea-mcp","status":"all"}`,
+			notQuery:  []string{"status"},
+			respBody:  `{"total_count":0,"workflow_runs":[]}`,
+			assert:    func(t *testing.T, out string) {},
+		},
+		{
+			name:      "branch filter",
+			input:     `{"owner":"mathias","name":"gitea-mcp","branch":"main"}`,
+			wantQuery: map[string]string{"branch": "main"},
+			respBody:  `{"total_count":0,"workflow_runs":[]}`,
+			assert:    func(t *testing.T, out string) {},
+		},
+	}
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+				assert.Equal(t, http.MethodGet, r.Method)
+				assert.Equal(t, "/api/v1/repos/mathias/gitea-mcp/actions/runs", r.URL.Path)
+				q := r.URL.Query()
+				for k, v := range tc.wantQuery {
+					assert.Equal(t, v, q.Get(k), "query param %q", k)
+				}
+				for _, k := range tc.notQuery {
+					assert.Equal(t, "", q.Get(k), "query param %q should be absent", k)
+				}
+				w.Header().Set("Content-Type", "application/json")
+				_, _ = w.Write([]byte(tc.respBody))
+			}))
+			defer srv.Close()
+
+			tool := tools.NewWorkflowRunList(gitea.NewClient(srv.URL, "tok"), allowlist.New([]string{"mathias"}))
+			out, err := tool.Call(context.Background(), json.RawMessage(tc.input))
+			require.NoError(t, err)
+			tc.assert(t, string(out))
+		})
+	}
+}
+
+func TestWorkflowRunListAllowlistRejects(t *testing.T) {
+	tool := tools.NewWorkflowRunList(gitea.NewClient("http://unused", ""), allowlist.New([]string{"mathias"}))
+	_, err := tool.Call(context.Background(), json.RawMessage(`{"owner":"evil","name":"x"}`))
+	require.Error(t, err)
+}