feat(auth): per-caller Gitea PAT pass-through (#59)

Replaces the shared GITEA_MCP_DEFAULT_TOKEN for all callers. When a
request's bearer validates directly against Gitea's own /api/v1/user,
that token is used for every upstream call this request makes instead
of the service PAT, and the caller identity comes from Gitea's own
login rather than the proxy header. Any other bearer (static token,
JWT, none) falls through unchanged to the existing chassis auth.

Prep: Authentik now SSOs into Gitea (infra a32801c), so each real user
can mint their own PAT from their own linked account.

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

.gitea/workflows/cd.yml

@@ -4,6 +4,10 @@ name: CD
   push:
     branches: [main]
     tags: ["v*"]
+    # Docs-only pushes don't change the image — skip the build/deploy roll.
+    # (Only applies to branch pushes; tag pushes always trigger.)
+    paths-ignore:
+      - 'docs/**'
 
 env:
   IMAGE: gitea-mcp

cmd/gitea-mcp/main.go

@@ -92,8 +92,10 @@ func main() {
 
 	mux := http.NewServeMux()
 	mux.Handle("/mcp", mcp.OriginAllowlist(cfg.OriginAllowlist)(
-		chassisauth.BearerMiddleware(cfg.StaticToken, jwtValidator, "gitea", resourceMetadataURL,
+		auth.PassthroughMiddleware(giteaClient, mcpSrv,
-			auth.CallerMiddleware(logger, mcpSrv),
+			chassisauth.BearerMiddleware(cfg.StaticToken, jwtValidator, "gitea", resourceMetadataURL,
+				auth.CallerMiddleware(logger, mcpSrv),
+			),
 		),
 	))
 	mux.Handle("/healthz", newHealthzHandler(cfg.DexIssuerURL != "", jwtValidator != nil, jwtInitErr))

internal/auth/caller.go

@@ -36,11 +36,14 @@ func CallerMiddleware(logger *slog.Logger, next http.Handler) http.Handler {
 				"x_forwarded_user", fwdUser)
 		}
 
-		ctx := context.WithValue(r.Context(), ctxKey{}, user)
+		next.ServeHTTP(w, r.WithContext(withCaller(r.Context(), user)))
-		next.ServeHTTP(w, r.WithContext(ctx))
 	})
 }
 
+func withCaller(ctx context.Context, user string) context.Context {
+	return context.WithValue(ctx, ctxKey{}, user)
+}
+
 func Caller(ctx context.Context) string {
 	if v, ok := ctx.Value(ctxKey{}).(string); ok {
 		return v

internal/auth/passthrough.go

@@ -0,0 +1,36 @@
+package auth
+
+import (
+	"context"
+	"net/http"
+	"strings"
+
+	"git.d-ma.be/mathias/gitea-mcp/internal/gitea"
+)
+
+// TokenValidator asks the upstream service who a bearer token belongs to.
+type TokenValidator interface {
+	ValidateToken(ctx context.Context, token string) (username string, ok bool)
+}
+
+// PassthroughMiddleware lets a caller authenticate with their own Gitea PAT:
+// if the request's bearer token validates directly against Gitea, it's used
+// as-is for every upstream call this request makes (gitea-mcp#59), instead of
+// the server's shared default token. Any other bearer (static token, JWT, or
+// none) falls through to fallback unchanged — this only adds a capability, it
+// never removes the existing auth paths.
+func PassthroughMiddleware(validator TokenValidator, onValid, fallback http.Handler) http.Handler {
+	return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		authz := r.Header.Get("Authorization")
+		token, hasBearer := strings.CutPrefix(authz, "Bearer ")
+		if hasBearer && token != "" {
+			if login, ok := validator.ValidateToken(r.Context(), token); ok {
+				ctx := withCaller(r.Context(), login)
+				ctx = gitea.WithToken(ctx, token)
+				onValid.ServeHTTP(w, r.WithContext(ctx))
+				return
+			}
+		}
+		fallback.ServeHTTP(w, r)
+	})
+}

internal/auth/passthrough_test.go

@@ -0,0 +1,82 @@
+package auth_test
+
+import (
+	"context"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"git.d-ma.be/mathias/gitea-mcp/internal/auth"
+	"git.d-ma.be/mathias/gitea-mcp/internal/gitea"
+	"github.com/stretchr/testify/assert"
+)
+
+type fakeValidator struct {
+	login string
+	ok    bool
+	calls int
+}
+
+func (f *fakeValidator) ValidateToken(_ context.Context, _ string) (string, bool) {
+	f.calls++
+	return f.login, f.ok
+}
+
+func TestPassthroughMiddleware_ValidPATGoesStraightToOnValid(t *testing.T) {
+	validator := &fakeValidator{login: "alice", ok: true}
+	var gotCaller string
+	var gotToken string
+	var gotOK bool
+	onValid := http.HandlerFunc(func(_ http.ResponseWriter, r *http.Request) {
+		gotCaller = auth.Caller(r.Context())
+		gotToken, gotOK = gitea.TokenFromContext(r.Context())
+	})
+	fallback := http.HandlerFunc(func(_ http.ResponseWriter, _ *http.Request) {
+		t.Fatal("fallback should not be called for a valid PAT")
+	})
+
+	h := auth.PassthroughMiddleware(validator, onValid, fallback)
+	req := httptest.NewRequest(http.MethodPost, "/mcp", nil)
+	req.Header.Set("Authorization", "Bearer alices-pat")
+	h.ServeHTTP(httptest.NewRecorder(), req)
+
+	assert.Equal(t, "alice", gotCaller)
+	assert.True(t, gotOK)
+	assert.Equal(t, "alices-pat", gotToken)
+}
+
+func TestPassthroughMiddleware_InvalidTokenFallsThrough(t *testing.T) {
+	validator := &fakeValidator{ok: false}
+	fallbackCalled := false
+	onValid := http.HandlerFunc(func(_ http.ResponseWriter, _ *http.Request) {
+		t.Fatal("onValid should not be called for an invalid token")
+	})
+	fallback := http.HandlerFunc(func(_ http.ResponseWriter, _ *http.Request) {
+		fallbackCalled = true
+	})
+
+	h := auth.PassthroughMiddleware(validator, onValid, fallback)
+	req := httptest.NewRequest(http.MethodPost, "/mcp", nil)
+	req.Header.Set("Authorization", "Bearer not-a-gitea-pat")
+	h.ServeHTTP(httptest.NewRecorder(), req)
+
+	assert.True(t, fallbackCalled)
+}
+
+func TestPassthroughMiddleware_NoBearerFallsThroughWithoutCallingValidator(t *testing.T) {
+	validator := &fakeValidator{ok: true, login: "alice"}
+	fallbackCalled := false
+	onValid := http.HandlerFunc(func(_ http.ResponseWriter, _ *http.Request) {
+		t.Fatal("onValid should not be called with no Authorization header")
+	})
+	fallback := http.HandlerFunc(func(_ http.ResponseWriter, _ *http.Request) {
+		fallbackCalled = true
+	})
+
+	h := auth.PassthroughMiddleware(validator, onValid, fallback)
+	req := httptest.NewRequest(http.MethodPost, "/mcp", nil)
+	h.ServeHTTP(httptest.NewRecorder(), req)
+
+	assert.True(t, fallbackCalled)
+	assert.Equal(t, 0, validator.calls, "validator should not be invoked when there's no bearer to check")
+}

internal/gitea/client.go

@@ -3,6 +3,7 @@ package gitea
 import (
 	"bytes"
 	"context"
+	"encoding/json"
 	"fmt"
 	"io"
 	"net/http"
@@ -19,6 +20,21 @@ type Client struct {
 	branchCache *expirable.LRU[string, string]
 }
 
+type ctxTokenKey struct{}
+
+// WithToken overrides the token used for upstream Gitea calls made with the
+// returned context, taking precedence over the Client's configured default
+// token. Used for per-caller PAT pass-through (gitea-mcp#59).
+func WithToken(ctx context.Context, token string) context.Context {
+	return context.WithValue(ctx, ctxTokenKey{}, token)
+}
+
+// TokenFromContext returns the token set by WithToken, if any.
+func TokenFromContext(ctx context.Context) (string, bool) {
+	v, ok := ctx.Value(ctxTokenKey{}).(string)
+	return v, ok
+}
+
 func NewClient(baseURL, token string) *Client {
 	return &Client{
 		baseURL:     baseURL,
@@ -28,6 +44,23 @@ func NewClient(baseURL, token string) *Client {
 	}
 }
 
+// ValidateToken asks Gitea who a given token belongs to (GET /api/v1/user
+// using that token, not the client's configured default token) and returns
+// its login name. Used for per-caller PAT pass-through (gitea-mcp#59).
+func (c *Client) ValidateToken(ctx context.Context, token string) (string, bool) {
+	body, status, err := c.doOnce(WithToken(ctx, token), http.MethodGet, "/api/v1/user", nil)
+	if err != nil || status != http.StatusOK {
+		return "", false
+	}
+	var user struct {
+		Login string `json:"login"`
+	}
+	if err := json.Unmarshal(body, &user); err != nil || user.Login == "" {
+		return "", false
+	}
+	return user.Login, true
+}
+
 // DefaultBranch returns the default branch for a repo. Cached for 60s.
 func (c *Client) DefaultBranch(ctx context.Context, owner, name string) (string, error) {
 	key := owner + "/" + name
@@ -66,6 +99,9 @@ func (c *Client) doOnce(ctx context.Context, method, path string, body []byte) (
 		return nil, 0, err
 	}
 	token := c.token
+	if override, ok := TokenFromContext(ctx); ok {
+		token = override
+	}
 	if token != "" {
 		req.Header.Set("Authorization", "token "+token)
 	}
@@ -135,6 +171,9 @@ func (c *Client) doRaw(ctx context.Context, method, path string, body []byte) (*
 		return nil, err
 	}
 	token := c.token
+	if override, ok := TokenFromContext(ctx); ok {
+		token = override
+	}
 	if token != "" {
 		req.Header.Set("Authorization", "token "+token)
 	}

internal/gitea/client_test.go

@@ -50,6 +50,49 @@ func TestRetryOn5xxGetSucceedsOnSecondAttempt(t *testing.T) {
 	assert.Equal(t, int32(2), atomic.LoadInt32(&attempts))
 }
 
+func TestClientPrefersTokenFromContextOverDefaultToken(t *testing.T) {
+	var gotAuth string
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		gotAuth = r.Header.Get("Authorization")
+		w.Header().Set("Content-Type", "application/json")
+		_, _ = w.Write([]byte(`{"ok":true}`))
+	}))
+	defer srv.Close()
+
+	c := gitea.NewClient(srv.URL, "default-token")
+	ctx := gitea.WithToken(context.Background(), "caller-token")
+	_, status, err := c.GetJSON(ctx, "/api/v1/user")
+	require.NoError(t, err)
+	assert.Equal(t, 200, status)
+	assert.Equal(t, "token caller-token", gotAuth)
+}
+
+func TestValidateTokenReturnsLoginOnSuccess(t *testing.T) {
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		assert.Equal(t, "token candidate-token", r.Header.Get("Authorization"))
+		w.Header().Set("Content-Type", "application/json")
+		_, _ = w.Write([]byte(`{"login":"alice"}`))
+	}))
+	defer srv.Close()
+
+	c := gitea.NewClient(srv.URL, "default-token")
+	login, ok := c.ValidateToken(context.Background(), "candidate-token")
+	assert.True(t, ok)
+	assert.Equal(t, "alice", login)
+}
+
+func TestValidateTokenReturnsFalseOn401(t *testing.T) {
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
+		http.Error(w, "unauthorized", http.StatusUnauthorized)
+	}))
+	defer srv.Close()
+
+	c := gitea.NewClient(srv.URL, "default-token")
+	login, ok := c.ValidateToken(context.Background(), "bad-token")
+	assert.False(t, ok)
+	assert.Empty(t, login)
+}
+
 func TestRetryOnPostNotRetried(t *testing.T) {
 	var attempts int32
 	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {

internal/gitea/code_search.go

@@ -1,10 +1,13 @@
 package gitea
 
 import (
+	"bytes"
 	"context"
-	"encoding/json"
+	"encoding/base64"
 	"fmt"
-	"net/url"
+	"path"
+	"sort"
+	"strings"
 )
 
 type CodeSearchHit struct {
@@ -14,30 +17,147 @@ type CodeSearchHit struct {
 	Score   float64 `json:"score,omitempty"`
 }
 
-type codeSearchEnvelope struct {
+// codeSearchMaxFiles bounds how many files a single SearchCode call will fetch
-	Data []CodeSearchHit `json:"data"`
+// and scan, so one call against a huge repo can't run indefinitely.
-	OK   bool            `json:"ok"`
+const codeSearchMaxFiles = 2000
+
+// codeSearchMaxFileSize skips blobs larger than this (per the tree listing,
+// before any fetch) — almost certainly binary/vendor/generated content, not
+// worth the cost of fetching just to reject.
+const codeSearchMaxFileSize = 512 * 1024
+
+// codeSearchBinaryExts are skipped WITHOUT fetching — a cheap pre-filter for
+// obviously-binary content by extension, checked against the tree listing.
+var codeSearchBinaryExts = map[string]bool{
+	".png": true, ".jpg": true, ".jpeg": true, ".gif": true, ".ico": true, ".webp": true, ".bmp": true,
+	".pdf": true, ".zip": true, ".tar": true, ".gz": true, ".bz2": true, ".xz": true, ".7z": true,
+	".exe": true, ".dll": true, ".so": true, ".dylib": true, ".bin": true, ".class": true, ".jar": true,
+	".woff": true, ".woff2": true, ".ttf": true, ".eot": true, ".otf": true,
+	".mp3": true, ".mp4": true, ".mov": true, ".avi": true, ".webm": true,
+	".pyc": true, ".o": true, ".a": true,
 }
 
+// SearchCode does a client-side "git grep"-equivalent search. Gitea's REST API
+// has no code-content-search endpoint — confirmed against a live 1.25.5
+// instance's swagger spec (only /repos/search, /repos/issues/search,
+// /topics/search etc exist). The web UI's OWN code search falls back to
+// server-side `git grep` because no Repository Indexer is enabled on that
+// instance, and that fallback is an HTML-only route, not JSON API. So this
+// walks the tree, fetches text-like blobs (bounded by codeSearchMaxFiles /
+// codeSearchMaxFileSize), and substring-matches q — case-insensitive, literal
+// (not a regex, to keep behavior predictable and avoid a ReDoS surface from
+// user-supplied input) — against file contents.
+//
+// Pagination is over the FULL sorted result set, recomputed on every call —
+// there is no server-side index to page through incrementally, so requesting
+// page 2 re-scans the tree. Acceptable for the repo sizes this targets; a real
+// indexer (bleve/elasticsearch) enabled server-side would be the long-term
+// fix, and is an infra decision, not something gitea-mcp controls.
 func (c *Client) SearchCode(ctx context.Context, owner, repo, q string, page, limit int) ([]CodeSearchHit, error) {
+	if q == "" {
+		return nil, fmt.Errorf("q is required: %w", ErrValidation)
+	}
 	if page < 1 {
 		page = 1
 	}
 	if limit < 1 {
 		limit = 30
 	}
-	path := fmt.Sprintf("/api/v1/repos/%s/%s/search?q=%s&type=code&page=%d&limit=%d",
+
-		owner, repo, url.QueryEscape(q), page, limit)
+	r, err := c.GetRepo(ctx, owner, repo)
-	body, status, err := c.GetJSON(ctx, path)
 	if err != nil {
-		return nil, err
+		return nil, fmt.Errorf("resolve repo: %w", err)
 	}
-	if err := MapStatus(status, body); err != nil {
+	branch := r.DefaultBranch
-		return nil, err
+	if branch == "" {
+		branch = "main"
 	}
-	var env codeSearchEnvelope
+
-	if err := json.Unmarshal(body, &env); err != nil {
+	tree, err := c.GetTree(ctx, owner, repo, branch, true)
-		return nil, err
+	if err != nil {
+		return nil, fmt.Errorf("tree walk: %w", err)
 	}
-	return env.Data, nil
+
+	qLower := strings.ToLower(q)
+	all := make([]CodeSearchHit, 0)
+	scanned := 0
+	for _, e := range tree.Tree {
+		if ctx.Err() != nil {
+			break
+		}
+		if e.Type != "blob" {
+			continue
+		}
+		if codeSearchBinaryExts[strings.ToLower(path.Ext(e.Path))] {
+			continue
+		}
+		if e.Size > codeSearchMaxFileSize {
+			continue
+		}
+		if scanned >= codeSearchMaxFiles {
+			break
+		}
+		scanned++
+
+		fc, ferr := c.GetFileContents(ctx, owner, repo, e.Path, branch)
+		if ferr != nil {
+			continue // vanished/unreadable between tree walk and read — skip, don't fail the whole search
+		}
+		decoded, derr := base64.StdEncoding.DecodeString(fc.Content)
+		if derr != nil {
+			continue
+		}
+		if bytes.IndexByte(decoded, 0) >= 0 {
+			continue // binary content the extension filter missed
+		}
+
+		content := string(decoded)
+		contentLower := strings.ToLower(content)
+		count := strings.Count(contentLower, qLower)
+		if count == 0 {
+			continue
+		}
+
+		idx := strings.Index(contentLower, qLower)
+		all = append(all, CodeSearchHit{
+			Path:    e.Path,
+			Snippet: codeSearchSnippet(content, idx, len(q)),
+			HTMLURL: fmt.Sprintf("%s/%s/%s/src/branch/%s/%s", c.baseURL, owner, repo, branch, e.Path),
+			Score:   float64(count),
+		})
+	}
+
+	sort.Slice(all, func(i, j int) bool {
+		if all[i].Score != all[j].Score {
+			return all[i].Score > all[j].Score
+		}
+		return all[i].Path < all[j].Path
+	})
+
+	start := (page - 1) * limit
+	if start >= len(all) {
+		return []CodeSearchHit{}, nil
+	}
+	end := start + limit
+	if end > len(all) {
+		end = len(all)
+	}
+	return all[start:end], nil
+}
+
+// codeSearchSnippet returns a short window of text centered on a match,
+// trimmed to a single line-ish window so results read like a grep hit rather
+// than a content dump.
+func codeSearchSnippet(content string, idx, matchLen int) string {
+	const window = 60
+	start := idx - window
+	if start < 0 {
+		start = 0
+	}
+	end := idx + matchLen + window
+	if end > len(content) {
+		end = len(content)
+	}
+	snippet := strings.ReplaceAll(content[start:end], "\n", " ")
+	return strings.TrimSpace(snippet)
 }

internal/gitea/code_search_test.go

@@ -2,8 +2,12 @@ package gitea_test
 
 import (
 	"context"
+	"encoding/base64"
+	"errors"
+	"fmt"
 	"net/http"
 	"net/http/httptest"
+	"strings"
 	"testing"
 
 	"git.d-ma.be/mathias/gitea-mcp/internal/gitea"
@@ -11,22 +15,65 @@ import (
 	"github.com/stretchr/testify/require"
 )
 
-func TestSearchCode(t *testing.T) {
+func b64(s string) string { return base64.StdEncoding.EncodeToString([]byte(s)) }
-	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+
-		assert.Equal(t, "/api/v1/repos/mathias/infra/search", r.URL.Path)
+// codeSearchFake serves GetRepo + GetTree + GetFileContents off an in-memory
-		assert.Equal(t, "SearchCode", r.URL.Query().Get("q"))
+// file map — the REAL endpoints SearchCode uses now that Gitea's REST API has
-		assert.Equal(t, "code", r.URL.Query().Get("type"))
+// no code-content-search endpoint (confirmed against a live 1.25.5 instance's
+// swagger spec: only /repos/search, /repos/issues/search etc exist — the web
+// UI's own code search falls back to server-side `git grep`, an HTML-only
+// route, not JSON API). Records which paths were actually fetched, so tests
+// can assert a file was SKIPPED (binary ext, oversized) without ever being
+// read, not just absent from results.
+type codeSearchFake struct {
+	files   map[string]string // path -> content
+	sizes   map[string]int64  // path -> tree-entry size (defaults to len(content))
+	fetched []string
+}
+
+func (f *codeSearchFake) handler(t *testing.T, owner, repo, branch string) http.HandlerFunc {
+	return func(w http.ResponseWriter, r *http.Request) {
 		w.Header().Set("Content-Type", "application/json")
-		_, _ = w.Write([]byte(`{
+		p := r.URL.Path
-			"data":[{
+		switch {
-				"path":"internal/gitea/code_search.go",
+		case r.Method == http.MethodGet && p == "/api/v1/repos/"+owner+"/"+repo:
-				"snippet":"func (c *Client) SearchCode",
+			_, _ = fmt.Fprintf(w, `{"name":%q,"full_name":"%s/%s","default_branch":%q}`, repo, owner, repo, branch)
-				"html_url":"http://gitea.example.com/mathias/infra/src/branch/main/internal/gitea/code_search.go",
+
-				"score":2.5
+		case r.Method == http.MethodGet && strings.HasPrefix(p, "/api/v1/repos/"+owner+"/"+repo+"/git/trees/"):
-			}],
+			var entries []string
-			"ok":true
+			for path, content := range f.files {
-		}`))
+				size := int64(len(content))
-	}))
+				if s, ok := f.sizes[path]; ok {
+					size = s
+				}
+				entries = append(entries, fmt.Sprintf(`{"path":%q,"type":"blob","sha":"s","size":%d}`, path, size))
+			}
+			_, _ = fmt.Fprintf(w, `{"sha":"root","tree":[%s],"truncated":false}`, strings.Join(entries, ","))
+
+		case r.Method == http.MethodGet && strings.HasPrefix(p, "/api/v1/repos/"+owner+"/"+repo+"/contents/"):
+			path := strings.TrimPrefix(p, "/api/v1/repos/"+owner+"/"+repo+"/contents/")
+			f.fetched = append(f.fetched, path)
+			content, ok := f.files[path]
+			if !ok {
+				w.WriteHeader(http.StatusNotFound)
+				_, _ = w.Write([]byte(`{"message":"not found"}`))
+				return
+			}
+			_, _ = fmt.Fprintf(w, `{"path":%q,"sha":"s","size":%d,"content":%q,"encoding":"base64"}`, path, len(content), b64(content))
+
+		default:
+			t.Errorf("unexpected request: %s %s", r.Method, p)
+			w.WriteHeader(http.StatusNotFound)
+		}
+	}
+}
+
+func TestSearchCode_FindsMatchInTree(t *testing.T) {
+	f := &codeSearchFake{files: map[string]string{
+		"internal/gitea/code_search.go": "func (c *Client) SearchCode(ctx context.Context) {}\n",
+		"internal/gitea/repos.go":       "func (c *Client) ListRepos() {}\n",
+	}}
+	srv := httptest.NewServer(f.handler(t, "mathias", "infra", "main"))
 	defer srv.Close()
 
 	c := gitea.NewClient(srv.URL, "tok")
@@ -34,6 +81,92 @@ func TestSearchCode(t *testing.T) {
 	require.NoError(t, err)
 	require.Len(t, hits, 1)
 	assert.Equal(t, "internal/gitea/code_search.go", hits[0].Path)
-	assert.Equal(t, "func (c *Client) SearchCode", hits[0].Snippet)
+	assert.Contains(t, hits[0].Snippet, "SearchCode")
-	assert.InDelta(t, 2.5, hits[0].Score, 0.001)
+	assert.Contains(t, hits[0].HTMLURL, "internal/gitea/code_search.go")
+	assert.Equal(t, 1.0, hits[0].Score)
+}
+
+func TestSearchCode_CaseInsensitive(t *testing.T) {
+	f := &codeSearchFake{files: map[string]string{
+		"README.md": "# MyProject\n\nBuild instructions here.\n",
+	}}
+	srv := httptest.NewServer(f.handler(t, "mathias", "infra", "main"))
+	defer srv.Close()
+
+	c := gitea.NewClient(srv.URL, "tok")
+	hits, err := c.SearchCode(context.Background(), "mathias", "infra", "myproject", 1, 30)
+	require.NoError(t, err)
+	require.Len(t, hits, 1)
+}
+
+func TestSearchCode_SkipsBinaryExtensionWithoutFetching(t *testing.T) {
+	f := &codeSearchFake{files: map[string]string{
+		"assets/logo.png": "SearchCode", // would match if fetched — must not be
+		"main.go":         "package main\n",
+	}}
+	srv := httptest.NewServer(f.handler(t, "mathias", "infra", "main"))
+	defer srv.Close()
+
+	c := gitea.NewClient(srv.URL, "tok")
+	hits, err := c.SearchCode(context.Background(), "mathias", "infra", "SearchCode", 1, 30)
+	require.NoError(t, err)
+	assert.Empty(t, hits)
+	assert.NotContains(t, f.fetched, "assets/logo.png", "binary extension must be skipped before fetch")
+}
+
+func TestSearchCode_SkipsOversizedFileWithoutFetching(t *testing.T) {
+	f := &codeSearchFake{
+		files: map[string]string{"vendor/bundle.txt": "SearchCode"},
+		sizes: map[string]int64{"vendor/bundle.txt": 10 * 1024 * 1024}, // 10MB per the tree listing
+	}
+	srv := httptest.NewServer(f.handler(t, "mathias", "infra", "main"))
+	defer srv.Close()
+
+	c := gitea.NewClient(srv.URL, "tok")
+	hits, err := c.SearchCode(context.Background(), "mathias", "infra", "SearchCode", 1, 30)
+	require.NoError(t, err)
+	assert.Empty(t, hits)
+	assert.NotContains(t, f.fetched, "vendor/bundle.txt", "oversized file must be skipped before fetch")
+}
+
+func TestSearchCode_SkipsBinaryContentNullByte(t *testing.T) {
+	f := &codeSearchFake{files: map[string]string{
+		"data.bin": "SearchCode\x00binary-marker",
+	}}
+	srv := httptest.NewServer(f.handler(t, "mathias", "infra", "main"))
+	defer srv.Close()
+
+	c := gitea.NewClient(srv.URL, "tok")
+	hits, err := c.SearchCode(context.Background(), "mathias", "infra", "SearchCode", 1, 30)
+	require.NoError(t, err)
+	assert.Empty(t, hits, "content with a null byte must be treated as binary even past the extension filter")
+}
+
+func TestSearchCode_Pagination(t *testing.T) {
+	files := map[string]string{}
+	for i := 0; i < 5; i++ {
+		files[fmt.Sprintf("file%d.go", i)] = strings.Repeat("hit ", i+1) // increasing occurrence count => increasing score
+	}
+	f := &codeSearchFake{files: files}
+	srv := httptest.NewServer(f.handler(t, "mathias", "infra", "main"))
+	defer srv.Close()
+
+	c := gitea.NewClient(srv.URL, "tok")
+	page1, err := c.SearchCode(context.Background(), "mathias", "infra", "hit", 1, 2)
+	require.NoError(t, err)
+	require.Len(t, page1, 2)
+	page2, err := c.SearchCode(context.Background(), "mathias", "infra", "hit", 2, 2)
+	require.NoError(t, err)
+	require.Len(t, page2, 2)
+
+	assert.NotEqual(t, page1[0].Path, page2[0].Path, "pages must not overlap")
+	assert.True(t, page1[0].Score >= page1[1].Score, "page1 sorted desc by score")
+	assert.True(t, page1[1].Score >= page2[0].Score, "page1's worst must rank >= page2's best")
+}
+
+func TestSearchCode_EmptyQueryErrors(t *testing.T) {
+	c := gitea.NewClient("http://unused", "tok")
+	_, err := c.SearchCode(context.Background(), "mathias", "infra", "", 1, 30)
+	require.Error(t, err)
+	assert.True(t, errors.Is(err, gitea.ErrValidation))
 }

internal/tools/code_search_test.go

@@ -2,8 +2,10 @@ package tools_test
 
 import (
 	"context"
+	"encoding/base64"
 	"encoding/json"
 	"errors"
+	"fmt"
 	"net/http"
 	"net/http/httptest"
 	"strings"
@@ -16,22 +18,76 @@ import (
 	"github.com/stretchr/testify/require"
 )
 
-func TestCodeSearchSingleRepo(t *testing.T) {
+// multiRepoSearchFake serves ListRepos + per-repo GetRepo/GetTree/GetFileContents
-	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+// for a set of repos — the real endpoints code_search's underlying SearchCode
-		assert.Equal(t, "/api/v1/repos/mathias/infra/search", r.URL.Path)
+// now uses (Gitea's REST API has no code-content-search endpoint; see
-		assert.Equal(t, "ListRepos", r.URL.Query().Get("q"))
+// internal/gitea/code_search.go's doc comment).
-		assert.Equal(t, "code", r.URL.Query().Get("type"))
+type multiRepoSearchFake struct {
+	owner string
+	repos []string                     // ListRepos response, in this order
+	files map[string]map[string]string // repo -> path -> content
+	fail  map[string]bool              // repo -> GetRepo 500s for this repo (simulates a per-repo failure)
+}
+
+func (f *multiRepoSearchFake) handler(t *testing.T) http.HandlerFunc {
+	return func(w http.ResponseWriter, r *http.Request) {
 		w.Header().Set("Content-Type", "application/json")
-		_, _ = w.Write([]byte(`{
+		p := r.URL.Path
-			"data":[{
+
-				"path":"internal/gitea/repos.go",
+		for _, repo := range f.repos {
-				"snippet":"func (c *Client) ListRepos",
+			base := "/api/v1/repos/" + f.owner + "/" + repo
-				"html_url":"http://gitea.example.com/mathias/infra/src/branch/main/internal/gitea/repos.go",
+			switch {
-				"score":3.0
+			case p == base && f.fail[repo]:
-			}],
+				w.WriteHeader(http.StatusInternalServerError)
-			"ok":true
+				_, _ = w.Write([]byte(`{"message":"internal error"}`))
-		}`))
+				return
-	}))
+			case p == base:
+				_, _ = fmt.Fprintf(w, `{"name":%q,"full_name":"%s/%s","default_branch":"main"}`, repo, f.owner, repo)
+				return
+			case strings.HasPrefix(p, base+"/git/trees/"):
+				var entries []string
+				for path := range f.files[repo] {
+					entries = append(entries, fmt.Sprintf(`{"path":%q,"type":"blob","sha":"s","size":100}`, path))
+				}
+				_, _ = fmt.Fprintf(w, `{"sha":"root","tree":[%s],"truncated":false}`, strings.Join(entries, ","))
+				return
+			case strings.HasPrefix(p, base+"/contents/"):
+				path := strings.TrimPrefix(p, base+"/contents/")
+				content, ok := f.files[repo][path]
+				if !ok {
+					w.WriteHeader(http.StatusNotFound)
+					_, _ = w.Write([]byte(`{"message":"not found"}`))
+					return
+				}
+				enc := base64.StdEncoding.EncodeToString([]byte(content))
+				_, _ = fmt.Fprintf(w, `{"path":%q,"sha":"s","size":%d,"content":%q,"encoding":"base64"}`, path, len(content), enc)
+				return
+			}
+		}
+
+		if p == "/api/v1/users/"+f.owner+"/repos" {
+			var entries []string
+			for _, repo := range f.repos {
+				entries = append(entries, fmt.Sprintf(`{"name":%q,"full_name":"%s/%s","default_branch":"main"}`, repo, f.owner, repo))
+			}
+			_, _ = fmt.Fprintf(w, `[%s]`, strings.Join(entries, ","))
+			return
+		}
+
+		t.Errorf("unexpected request: %s %s", r.Method, p)
+		w.WriteHeader(http.StatusNotFound)
+	}
+}
+
+func TestCodeSearchSingleRepo(t *testing.T) {
+	f := &multiRepoSearchFake{
+		owner: "mathias",
+		repos: []string{"infra"},
+		files: map[string]map[string]string{
+			"infra": {"internal/gitea/repos.go": "func (c *Client) ListRepos() {}\n"},
+		},
+	}
+	srv := httptest.NewServer(f.handler(t))
 	defer srv.Close()
 
 	tool := tools.NewCodeSearch(gitea.NewClient(srv.URL, "tok"), allowlist.New([]string{"mathias"}))
@@ -50,7 +106,7 @@ func TestCodeSearchSingleRepo(t *testing.T) {
 	require.Len(t, result.Results, 1)
 	assert.Equal(t, "mathias/infra", result.Results[0].Repo)
 	assert.Equal(t, "internal/gitea/repos.go", result.Results[0].Path)
-	assert.Equal(t, "func (c *Client) ListRepos", result.Results[0].Snippet)
+	assert.Contains(t, result.Results[0].Snippet, "ListRepos")
 }
 
 func TestCodeSearchAllowlistRejects(t *testing.T) {
@@ -67,22 +123,15 @@ func TestCodeSearchRequiresQ(t *testing.T) {
 }
 
 func TestCodeSearchFanOutHappyPath(t *testing.T) {
-	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+	f := &multiRepoSearchFake{
-		w.Header().Set("Content-Type", "application/json")
+		owner: "mathias",
-		switch r.URL.Path {
+		repos: []string{"infra", "gitea-mcp"},
-		case "/api/v1/users/mathias/repos":
+		files: map[string]map[string]string{
-			_, _ = w.Write([]byte(`[
+			"infra":     {"main.go": "this is an infra hit\n"},
-				{"name":"infra","full_name":"mathias/infra","default_branch":"main"},
+			"gitea-mcp": {"cmd/main.go": "this is a gitea-mcp hit\n"},
-				{"name":"gitea-mcp","full_name":"mathias/gitea-mcp","default_branch":"main"}
+		},
-			]`))
+	}
-		case "/api/v1/repos/mathias/infra/search":
+	srv := httptest.NewServer(f.handler(t))
-			_, _ = w.Write([]byte(`{"data":[{"path":"main.go","snippet":"infra hit","html_url":"http://x/infra/main.go","score":2.0}],"ok":true}`))
-		case "/api/v1/repos/mathias/gitea-mcp/search":
-			_, _ = w.Write([]byte(`{"data":[{"path":"cmd/main.go","snippet":"gitea-mcp hit","html_url":"http://x/gitea-mcp/main.go","score":1.0}],"ok":true}`))
-		default:
-			http.NotFound(w, r)
-		}
-	}))
 	defer srv.Close()
 
 	tool := tools.NewCodeSearch(gitea.NewClient(srv.URL, "tok"), allowlist.New([]string{"mathias"}))
@@ -110,23 +159,15 @@ func TestCodeSearchFanOutHappyPath(t *testing.T) {
 }
 
 func TestCodeSearchFanOutPartialFailure(t *testing.T) {
-	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+	f := &multiRepoSearchFake{
-		w.Header().Set("Content-Type", "application/json")
+		owner: "mathias",
-		switch r.URL.Path {
+		repos: []string{"infra", "broken"},
-		case "/api/v1/users/mathias/repos":
+		files: map[string]map[string]string{
-			_, _ = w.Write([]byte(`[
+			"infra": {"main.go": "this is an infra hit\n"},
-				{"name":"infra","full_name":"mathias/infra","default_branch":"main"},
+		},
-				{"name":"broken","full_name":"mathias/broken","default_branch":"main"}
+		fail: map[string]bool{"broken": true},
-			]`))
+	}
-		case "/api/v1/repos/mathias/infra/search":
+	srv := httptest.NewServer(f.handler(t))
-			_, _ = w.Write([]byte(`{"data":[{"path":"main.go","snippet":"infra hit","html_url":"http://x/infra/main.go","score":1.0}],"ok":true}`))
-		case "/api/v1/repos/mathias/broken/search":
-			w.WriteHeader(http.StatusInternalServerError)
-			_, _ = w.Write([]byte(`{"message":"internal error"}`))
-		default:
-			http.NotFound(w, r)
-		}
-	}))
 	defer srv.Close()
 
 	tool := tools.NewCodeSearch(gitea.NewClient(srv.URL, "tok"), allowlist.New([]string{"mathias"}))
@@ -134,9 +175,11 @@ func TestCodeSearchFanOutPartialFailure(t *testing.T) {
 	require.NoError(t, err)
 
 	var result struct {
-		Results      []struct{ Repo string `json:"repo"` } `json:"results"`
+		Results []struct {
-		Partial      bool                                  `json:"partial"`
+			Repo string `json:"repo"`
-		PartialRepos []string                              `json:"partial_repos"`
+		} `json:"results"`
+		Partial      bool     `json:"partial"`
+		PartialRepos []string `json:"partial_repos"`
 	}
 	require.NoError(t, json.Unmarshal(out, &result))
 	assert.True(t, result.Partial)
@@ -147,41 +190,31 @@ func TestCodeSearchFanOutPartialFailure(t *testing.T) {
 }
 
 func TestCodeSearchFanOutSortsByScore(t *testing.T) {
-	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+	f := &multiRepoSearchFake{
-		w.Header().Set("Content-Type", "application/json")
+		owner: "mathias",
-		switch r.URL.Path {
+		repos: []string{"alpha", "beta"},
-		case "/api/v1/users/mathias/repos":
+		files: map[string]map[string]string{
-			_, _ = w.Write([]byte(`[
+			"alpha": {"a.go": "one high here"},            // 1 occurrence => score 1
-				{"name":"alpha","full_name":"mathias/alpha","default_branch":"main"},
+			"beta":  {"b.go": "high high high high high"}, // 5 occurrences => score 5
-				{"name":"beta","full_name":"mathias/beta","default_branch":"main"}
+		},
-			]`))
+	}
-		case "/api/v1/repos/mathias/alpha/search":
+	srv := httptest.NewServer(f.handler(t))
-			// low score
-			_, _ = w.Write([]byte(`{"data":[{"path":"a.go","snippet":"low","html_url":"http://x/alpha/a.go","score":1.0}],"ok":true}`))
-		case "/api/v1/repos/mathias/beta/search":
-			// high score
-			_, _ = w.Write([]byte(`{"data":[{"path":"b.go","snippet":"high","html_url":"http://x/beta/b.go","score":5.0}],"ok":true}`))
-		default:
-			http.NotFound(w, r)
-		}
-	}))
 	defer srv.Close()
 
 	tool := tools.NewCodeSearch(gitea.NewClient(srv.URL, "tok"), allowlist.New([]string{"mathias"}))
-	out, err := tool.Call(context.Background(), json.RawMessage(`{"q":"something","owner":"mathias"}`))
+	out, err := tool.Call(context.Background(), json.RawMessage(`{"q":"high","owner":"mathias"}`))
 	require.NoError(t, err)
 
 	var result struct {
 		Results []struct {
+			Repo    string  `json:"repo"`
 			Snippet string  `json:"snippet"`
 			Score   float64 `json:"score"`
 		} `json:"results"`
 	}
 	require.NoError(t, json.Unmarshal(out, &result))
 	require.Len(t, result.Results, 2)
-	// First result must be the high-score one
+	assert.Equal(t, "mathias/beta", result.Results[0].Repo, "higher-score repo (5 occurrences) must sort first")
 	assert.True(t, result.Results[0].Score > result.Results[1].Score,
-		"expected results sorted by score desc, got %v then %v",
+		"expected results sorted by score desc, got %v then %v", result.Results[0].Score, result.Results[1].Score)
-		result.Results[0].Score, result.Results[1].Score)
-	assert.True(t, strings.Contains(result.Results[0].Snippet, "high"))
 }