feat(stuck): stuck-state pipeline overlay (cad-atlas#11)

Consumes mathias/cad#14's cad-stuck query directly — cad's app/domain
packages are public Go modules, so this reuses the exact tested
classification logic (import, not re-implementation) rather than
duplicating the taxonomy across two repos.

ApplyStuckOverlay (internal/atlas/stuck.go, fixture-tested) maps each
finding's category onto the existing 9-stage pipeline: armed_never_fired to
the human dispatch gate, claimed_never_ran/ceiling_hit/refire_loop/
near_attempt_cap to the one Execute stage (there is no separate Claim
stage), done_not_closed to the loop-back stage. Adds a StuckCount/
AwaitingCount badge per stage plus one Node per finding, linking out to the
underlying issue/comment — read-only, no controls, matching cad#14's own
constraint.

Two visual classes, kept deliberately distinct per the issue's own framing:
a coral 'N stuck' badge (needs intervention) vs a green 'N awaiting' badge
(a functioning control waiting on a human) — done_not_closed can never
render as a bottleneck.

Verified against the live estate: STAGE 05 shows 2 stuck (cobalt-dingo#30
ceiling_hit, cad-atlas#10 near_attempt_cap), STAGE 08 shows 5 awaiting
(cobalt-dingo #3/#7/#12/#16/#19 done_not_closed) — screenshotted via a
koala k3s Playwright job against a live-token server, matching cad#14's own
proven query output exactly.

go.mod: added git.d-ma.be/mathias/cad v0.10.0 (app/domain/adapters-gitearead
only — no mcp-chassis pulled in, that's only imported by cad's own cmd/cad).
Dockerfile: added the same GOPRIVATE/insteadOf fetch config mathias/cad's
own Dockerfile uses, now needed for this new private-module dependency.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

internal/atlas/atlas.json

@@ -29,12 +29,12 @@
       {"t":"Admission controller","plain_t":"Tamper-proof seal","plain":"The work order is cryptographically signed when created, so any later tampering is detectable and the eventual change can be checked against it.","d":"Ed25519-sign issue body at creation (#36). Verify sig + PR alignment at infra boundary.","tags":["chain of custody"]},
       {"t":"⚖️ var-go Oath","plain_t":"Definition of done","plain":"A machine-checkable \"definition of done\" is embedded in the work order — exactly one, or the order is rejected — later used to prove the result actually meets the spec.","cls":"oath","pill":"var(--gold)","d":"Acceptance contract embedded in the issue as a var fenced block. Exactly one — zero/multiple fail closed. Prose → typed steps; failures anchored to byte spans.","tags":["swedsl · var-go","defined here → enforced @06"]}
     ]},
-    {"no":"STAGE 04","short":"Human go","cls":"gate","title":"Human dispatch gate","plain_title":"Human says go","plain":"A person reviews the work order and its risk and decides whether to release it — the one and only checkpoint where work does not move on its own.","path":"the only checkpoint",
+    {"no":"STAGE 04","short":"Human go","cls":"gate","title":"Human dispatch gate","plain_title":"Human says go","plain":"A person reviews the work order and its risk and decides whether to release it — the one and only checkpoint where work does not move on its own.","path":"the only checkpoint","generate":"stuck-badges",
      "trans_label":"A human said go","trans":"The hard stop. Nothing crosses automatically — a person must review the plan and risk and explicitly release it, and the repo must be on the allow-list, before any agent starts. This is the single human checkpoint in the whole pipeline.","nodes":[
       {"t":"Human triggers execution","plain_t":"The go button","plain":"A person confirms the plan and its risk level, then releases the work — nothing runs until they do.","cls":"gateway","pill":"var(--amber)","d":"Ratify proposed-plan + risk tier, then dispatch.","gate":true},
       {"t":"Session-Dispatch bridge","plain_t":"Hand-off to agents","plain":"The approval flips a switch that hands the signed work order over to the agents to start execution.","cls":"bridge","pill":"var(--blue)","d":"claude.ai MCP → gitea:workflow_run_trigger → cad-dispatch.yml → agentsquad. The final design→execution bridge.","tags":["workflow_dispatch"]}
     ]},
-    {"no":"STAGE 05","short":"Build","cls":"exec","title":"Execute · agentsquad + dispatch","plain_title":"Agents do the work","plain":"AI agents actually build the thing — one writes, a second independent one reviews it to avoid marking its own homework — and every step is logged for the audit trail. Two execution paths run today: agentsquad for orchestrated multi-agent jobs, and dispatch for headless single-issue work.","path":"koala · agentsquad-serve + dispatch CronJob",
+    {"no":"STAGE 05","short":"Build","cls":"exec","title":"Execute · agentsquad + dispatch","plain_title":"Agents do the work","plain":"AI agents actually build the thing — one writes, a second independent one reviews it to avoid marking its own homework — and every step is logged for the audit trail. Two execution paths run today: agentsquad for orchestrated multi-agent jobs, and dispatch for headless single-issue work.","path":"koala · agentsquad-serve + dispatch CronJob","generate":"stuck-badges",
      "trans_label":"Change proposed","trans":"Advances when the agents finish and open a proposed change (a PR) with its audit log attached. Until there's a concrete change to test, nothing moves.","nodes":[
       {"t":"Task API","plain_t":"Start a job","plain":"A request kicks off a job and hands back an id you can poll for progress.","pill":"var(--coral)","d":"POST /tasks → job id · GET /tasks/{id}. taskqueue + serve (v0.12+).","tags":["single agentsquad.yaml"]},
       {"t":"Executor + reviewer loop","plain_t":"Build + independent review","plain":"One agent does the work; a second, independent agent on a different model reviews it — so nothing marks its own homework.","cls":"win","pill":"var(--coral)","d":"ADK Go + LiteLLM. Frontier models (local qwen spirals). Reviewer on distinct tier — echo-chamber prevention.","risk":true},
@@ -53,7 +53,7 @@
      "trans_label":"It's live","trans":"Once the new version is actually running on the server, the deployed outcome becomes the input to scoring. Advancing means shipped and observable, not just merged.","nodes":[
       {"t":"Deploy on green","plain_t":"Auto-deploy when green","plain":"When all checks pass, the release system rolls the new version onto the live server automatically — but only once the release is pointed at that version (merging code alone doesn't ship it).","pill":"var(--green)","d":"Flux reconciles image → k3s pod on koala. Push ≠ deploy: bump tag in mathias/infra.","tags":["ntfy on deploy"]}
     ]},
-    {"no":"STAGE 08","short":"Learn","cls":"telos","title":"Loop back","plain_title":"Did it work?","plain":"The result is scored against the goal that started it and fed back into the mission board, so the next round of planning learns from what shipped.","path":"→ TELOS (feedback bus)",
+    {"no":"STAGE 08","short":"Learn","cls":"telos","title":"Loop back","plain_title":"Did it work?","plain":"The result is scored against the goal that started it and fed back into the mission board, so the next round of planning learns from what shipped.","path":"→ TELOS (feedback bus)","generate":"stuck-badges",
      "trans_label":"What did we learn?","trans":"The scored outcome flows back into the mission board so goals, problems, and priorities update — the loop that makes the pipeline a cycle rather than a line. Partly manual today; an explicit improvement target.","nodes":[
       {"t":"Close the loop","plain_t":"Score & feed back","plain":"The outcome is scored against the goal that started it and written back to the mission board, so future planning learns from what actually shipped.","pill":"var(--violet)","d":"session_log + attestation → brain. Score deploy outcome vs originating goal. (arc partly manual — improvement target.)","tags":["continuous"]}
     ]}

internal/atlas/model.go

@@ -42,7 +42,13 @@ type Stage struct {
 	Generate   string `json:"generate,omitempty"`
 	TransLabel string `json:"trans_label,omitempty"`
 	Trans      string `json:"trans,omitempty"`
-	Nodes      []Node `json:"nodes"`
+	// StuckCount/AwaitingCount are set by ApplyStuckOverlay (cad-atlas#11) on
+	// stages marked `"generate":"stuck-badges"` — the two visual classes the
+	// issue asks for kept as separate counts so a functioning control
+	// (awaiting human) can never be rendered as a bottleneck (stuck).
+	StuckCount    int    `json:"stuck_count,omitempty"`
+	AwaitingCount int    `json:"awaiting_count,omitempty"`
+	Nodes         []Node `json:"nodes"`
 }
 
 // Atlas is the full data model the frontend renders.

internal/atlas/stuck.go

@@ -0,0 +1,100 @@
+// stuck.go overlays cad-atlas#11's stuck-state badges onto the pipeline: a
+// count badge per stage (Node entries + Stage.StuckCount/AwaitingCount),
+// sourced from mathias/cad#14's `cad stuck` query rather than re-implemented
+// here. cad's app/domain packages are public Go modules (not internal/) —
+// this reuses the exact tested classification logic instead of duplicating
+// the taxonomy, per ADR-0002 ("a small coupling surface, not a monolith").
+package atlas
+
+import (
+	"fmt"
+
+	caddomain "git.d-ma.be/mathias/cad/domain"
+)
+
+// stageForCategory maps a cad-stuck taxonomy row onto the existing 9-stage
+// pipeline (atlas.json), following the grouping cad-atlas#11 itself proposes:
+// Trigger→armed-but-never-fired sits at the human dispatch gate (the label
+// that authorizes dispatch is on, nothing has claimed it yet); Claim/Execute
+// collapse into the one Execute stage this pipeline already has (there is no
+// separate Claim stage); Close maps to the loop-back stage, where a human
+// closing the issue is the action that actually completes the loop.
+//
+// CategoryDoneNotClosed is deliberately absent from the "stuck" class — see
+// ApplyStuckOverlay, which reads Class rather than Category to decide the
+// visual treatment. It IS present here for stage placement.
+var stageForCategory = map[caddomain.StuckCategory]string{
+	caddomain.CategoryArmedNeverFired: "STAGE 04",
+	caddomain.CategoryClaimedNeverRan: "STAGE 05",
+	caddomain.CategoryCeilingHit:      "STAGE 05",
+	caddomain.CategoryReFireLoop:      "STAGE 05",
+	caddomain.CategoryNearAttemptCap:  "STAGE 05",
+	caddomain.CategoryDoneNotClosed:   "STAGE 08",
+}
+
+// stuckNodeCls / awaitingNodeCls are the CSS classes the static shell colours
+// (cad-atlas.html: .node.stuck, .node.awaiting-human) — the two visual
+// classes cad-atlas#11 asks for, kept visually distinct so "awaiting human"
+// (a functioning control) never reads as "stuck" (needs intervention).
+const (
+	stuckNodeCls    = "stuck"
+	awaitingNodeCls = "awaiting-human"
+)
+
+// ApplyStuckOverlay assigns each finding to the stage stageForCategory names,
+// and for every stage in stages marked `"generate":"stuck-badges"`, sets its
+// StuckCount/AwaitingCount and prepends one Node per finding assigned there
+// (title, class, repo tag, and a URL to the finding's first locator — this
+// overlay never adds a control, only a link out to Gitea). Findings whose
+// category has no stage mapping are silently omitted from the pipeline view
+// rather than misplaced — cad-atlas#11 is a read of the query, not a second
+// copy of its taxonomy, so an unmapped category is a gap to notice in
+// stageForCategory, not something to guess a stage for.
+//
+// stages is returned unmodified aside from the stuck-badges stages; this is a
+// pure function, no I/O, so it is fixture-testable without a live query.
+func ApplyStuckOverlay(stages []Stage, findings []caddomain.StuckFinding) []Stage {
+	byStage := make(map[string][]caddomain.StuckFinding)
+	for _, f := range findings {
+		stageNo, ok := stageForCategory[f.Category]
+		if !ok {
+			continue
+		}
+		byStage[stageNo] = append(byStage[stageNo], f)
+	}
+
+	out := make([]Stage, len(stages))
+	copy(out, stages)
+	for i := range out {
+		if out[i].Generate != "stuck-badges" {
+			continue
+		}
+		assigned := byStage[out[i].No]
+		var stuckN, awaitingN int
+		nodes := make([]Node, 0, len(assigned))
+		for _, f := range assigned {
+			cls := stuckNodeCls
+			if f.Class == caddomain.ClassAwaitingHuman {
+				cls = awaitingNodeCls
+				awaitingN++
+			} else {
+				stuckN++
+			}
+			var url string
+			if len(f.Locators) > 0 {
+				url = f.Locators[0]
+			}
+			nodes = append(nodes, Node{
+				Title: fmt.Sprintf("#%d %s", f.Issue, f.Category),
+				Desc:  f.Detail,
+				Cls:   cls,
+				Tags:  []string{"live · cad-stuck", f.Repo},
+				URL:   url,
+			})
+		}
+		out[i].StuckCount = stuckN
+		out[i].AwaitingCount = awaitingN
+		out[i].Nodes = append(nodes, out[i].Nodes...)
+	}
+	return out
+}

internal/atlas/stuck_test.go

@@ -0,0 +1,110 @@
+package atlas_test
+
+import (
+	"testing"
+
+	"git.d-ma.be/mathias/cad-atlas/internal/atlas"
+	caddomain "git.d-ma.be/mathias/cad/domain"
+)
+
+func stage(no, generate string) atlas.Stage {
+	return atlas.Stage{No: no, Title: "example", Generate: generate}
+}
+
+func finding(repo string, issue int, cat caddomain.StuckCategory, class caddomain.StuckClass) caddomain.StuckFinding {
+	return caddomain.StuckFinding{
+		Repo: repo, Issue: issue, Title: "example", Category: cat, Class: class,
+		Detail:   "detail",
+		Locators: []string{"https://git.d-ma.be/" + repo + "/issues/1"},
+	}
+}
+
+func TestApplyStuckOverlay_BadgesOnlyOnGeneratedStages(t *testing.T) {
+	stages := []atlas.Stage{
+		stage("STAGE 04", ""),              // no generate marker — must be untouched
+		stage("STAGE 04b", "stuck-badges"), // marked but no findings map here
+	}
+	out := atlas.ApplyStuckOverlay(stages, nil)
+	if len(out[0].Nodes) != 0 || out[0].StuckCount != 0 {
+		t.Fatalf("un-marked stage was touched: %+v", out[0])
+	}
+}
+
+func TestApplyStuckOverlay_ArmedNeverFiredGoesToStage04(t *testing.T) {
+	stages := []atlas.Stage{stage("STAGE 04", "stuck-badges")}
+	findings := []caddomain.StuckFinding{
+		finding("mathias/brain-gardener", 190, caddomain.CategoryArmedNeverFired, caddomain.ClassStuck),
+	}
+	out := atlas.ApplyStuckOverlay(stages, findings)
+	if out[0].StuckCount != 1 || out[0].AwaitingCount != 0 {
+		t.Fatalf("counts = stuck=%d awaiting=%d, want stuck=1 awaiting=0", out[0].StuckCount, out[0].AwaitingCount)
+	}
+	if len(out[0].Nodes) != 1 {
+		t.Fatalf("want 1 node, got %d", len(out[0].Nodes))
+	}
+	n := out[0].Nodes[0]
+	if n.Cls != "stuck" {
+		t.Errorf("Cls = %q, want %q", n.Cls, "stuck")
+	}
+	if n.URL == "" {
+		t.Error("node has no URL locator")
+	}
+	if n.Title != "#190 armed_never_fired" {
+		t.Errorf("Title = %q", n.Title)
+	}
+}
+
+func TestApplyStuckOverlay_DoneNotClosedIsAwaitingHumanAtStage08(t *testing.T) {
+	stages := []atlas.Stage{stage("STAGE 08", "stuck-badges")}
+	findings := []caddomain.StuckFinding{
+		finding("mathias/cobalt-dingo", 19, caddomain.CategoryDoneNotClosed, caddomain.ClassAwaitingHuman),
+		finding("mathias/cobalt-dingo", 16, caddomain.CategoryDoneNotClosed, caddomain.ClassAwaitingHuman),
+	}
+	out := atlas.ApplyStuckOverlay(stages, findings)
+	if out[0].StuckCount != 0 || out[0].AwaitingCount != 2 {
+		t.Fatalf("counts = stuck=%d awaiting=%d, want stuck=0 awaiting=2 — done_not_closed must never render as stuck", out[0].StuckCount, out[0].AwaitingCount)
+	}
+	for _, n := range out[0].Nodes {
+		if n.Cls != "awaiting-human" {
+			t.Errorf("node Cls = %q, want %q", n.Cls, "awaiting-human")
+		}
+	}
+}
+
+func TestApplyStuckOverlay_ExecutionCategoriesCollapseToStage05(t *testing.T) {
+	stages := []atlas.Stage{stage("STAGE 05", "stuck-badges")}
+	findings := []caddomain.StuckFinding{
+		finding("mathias/cobalt-dingo", 30, caddomain.CategoryCeilingHit, caddomain.ClassStuck),
+		finding("mathias/cad-atlas", 10, caddomain.CategoryNearAttemptCap, caddomain.ClassStuck),
+		finding("mathias/cobalt-dingo", 16, caddomain.CategoryReFireLoop, caddomain.ClassStuck),
+		finding("mathias/cobalt-dingo", 41, caddomain.CategoryClaimedNeverRan, caddomain.ClassStuck),
+	}
+	out := atlas.ApplyStuckOverlay(stages, findings)
+	if out[0].StuckCount != 4 {
+		t.Fatalf("StuckCount = %d, want 4", out[0].StuckCount)
+	}
+}
+
+func TestApplyStuckOverlay_NodesPrependedNotReplacingAuthoredContent(t *testing.T) {
+	s := stage("STAGE 04", "stuck-badges")
+	s.Nodes = []atlas.Node{{Title: "authored node"}}
+	out := atlas.ApplyStuckOverlay([]atlas.Stage{s}, []caddomain.StuckFinding{
+		finding("mathias/brain-gardener", 190, caddomain.CategoryArmedNeverFired, caddomain.ClassStuck),
+	})
+	if len(out[0].Nodes) != 2 {
+		t.Fatalf("want 2 nodes (1 live + 1 authored), got %d", len(out[0].Nodes))
+	}
+	if out[0].Nodes[len(out[0].Nodes)-1].Title != "authored node" {
+		t.Errorf("authored node was not preserved: %+v", out[0].Nodes)
+	}
+}
+
+func TestApplyStuckOverlay_DoesNotMutateInput(t *testing.T) {
+	stages := []atlas.Stage{stage("STAGE 04", "stuck-badges")}
+	_ = atlas.ApplyStuckOverlay(stages, []caddomain.StuckFinding{
+		finding("mathias/brain-gardener", 190, caddomain.CategoryArmedNeverFired, caddomain.ClassStuck),
+	})
+	if len(stages[0].Nodes) != 0 {
+		t.Fatalf("input slice was mutated: %+v", stages[0].Nodes)
+	}
+}