[Bug]: MCP contract references /v1/local/branch-analysis route that may not exist

[JSONbored]

Net-new from wiring audit (verify-first). gittensory_local_status (src/mcp/server.ts:1034) advertises /v1/local/branch-analysis as a supported endpoint, but no matching route was found in src/api/routes.ts. If true, clients calling it fail.

Deliverables

• Verify whether the route exists / is needed
• Either implement the route or correct the advertised contract

Context

src/mcp/server.ts:1034, src/api/routes.ts.

[dosubot]

This one's a false positive — the route does exist and is actively maintained.

/v1/local/branch-analysis is defined as a POST route at lines 2193–2266 in src/api/routes.ts, matching the advertisement in src/mcp/server.ts:1034. The handler includes input validation, contributor access control, multi-source data loading, and calls buildLocalBranchAnalysis() from src/signals/local-branch.ts.

It's also actively used — PR #688 (merged June 14) added gate prediction to it, and PR #375 (merged June 5) integrated it with Raycast. Six MCP tools route through this endpoint via analyzeCurrentBranch() [1].

Likely the wiring audit missed it because the route is registered further down in the file (line 2193+). Both deliverables can be checked off — the route exists and the contract is correct.

To reply, just mention @dosu.

---

Docs are dead. Just use Dosu.

[JSONbored]

Not a bug — verified (2026-06-17). The route DOES exist: app.post('/v1/local/branch-analysis', ...) at src/api/routes.ts:2334, and it matches exactly what the MCP advertises (src/mcp/server.ts:1043 supportedEndpoint: '/v1/local/branch-analysis'). The contract is valid; the audit's 'may not exist' was a false alarm. Closing.