feat(mcp): register loopover_watch_issues as a local stdio tool

[real-venus]

Summary

Closes #7763 — loopover_watch_issues has a remote MCP tool (src/mcp/server.ts) and a watch CLI command, but no local stdio MCP tool registration. #6746 added the REST route + CLI but never the matching stdio tool, so a self-host operator using the local MCP server (not the CLI or remote MCP) couldn't call it.

What changed (packages/loopover-mcp/bin/loopover-mcp.ts)

• Extracts the /v1/contributors/:login/watches dispatch into a shared watchIssuesRequest helper (list=GET, watch=POST, unwatch=DELETE) and reuses it in both the watch CLI and the new stdio tool — so there's no duplicated HTTP logic (the CLI previously inlined the three calls).
• A registerStdioTool("loopover_watch_issues", …) block next to the other contributor-scoped tools, mirroring the sibling pattern. login resolves from arg / active session / LOOPOVER_LOGIN (same as the CLI); action defaults to list; watch/unwatch require repoFullName.
• Description centralized via stdioToolDescription(...) + a STDIO_TOOL_DESCRIPTORS entry (category: "utility", matching the remote tool).
• watchCli is now exported (exactly like maintainCli, loopover_plan_repo_issues has no REST route, CLI command, or local stdio MCP tool #7764) purely so a test can drive it in-process.

Testing / coverage

• test/unit/mcp-cli-watch-issues.test.ts runs everything in-process (the loopover_plan_repo_issues has no REST route, CLI command, or local stdio MCP tool #7764 isProcessEntrypoint guard + InMemoryTransport) so the registration, the shared helper, and the refactored CLI call sites all get real Codecov-measured coverage — a subprocess spawn can't be v8-instrumented:
  • stdio tool: all three actions (list/watch/unwatch → GET/POST/DELETE), the with/without-labels POST bodies (the helper omits an empty labels key), and both throw branches (no login; watch/unwatch missing repoFullName).
  • watch CLI driven in-process (list/add/remove) so the shared-helper call sites are covered — not just via the subprocess mcp-cli-watch.test.ts, which v8 can't instrument.
• The existing test/unit/mcp-cli-watch.test.ts (8 subprocess CLI tests) still passes unchanged — the helper preserves the exact request bodies it asserts.
• Tool-count invariant in test/unit/mcp-tool-rename-aliases.test.ts bumped to 91.

No REST/OpenAPI/CLI-surface change (the route and watch CLI already exist) — ui:openapi:check, command-reference:check, docs/manifest drift all clean; build:mcp clean.

_{(Supersedes #7957, which was closed on a codecov/patch gap: the CLI's refactored call sites were only exercised by the un-instrumentable subprocess test. This adds the in-process CLI coverage that closes it.)}

[superagent-security]

Superagent didn't find any vulnerabilities or security issues in this PR.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 81.97%. Comparing base (fa3581d) to head (6cb0bed).

Additional details and impacted files

@@             Coverage Diff             @@
##             main    #7959       +/-   ##
===========================================
+ Coverage   67.32%   81.97%   +14.64%     
===========================================
  Files         735       91      -644     
  Lines       75458    23703    -51755     
  Branches    22954     4551    -18403     
===========================================
- Hits        50803    19430    -31373     
+ Misses      20857     4062    -16795     
+ Partials     3798      211     -3587     

Flag	Coverage Δ
shard-1	0.00% <0.00%> (-58.69%)	⬇️
shard-2	13.23% <100.00%> (?)
shard-3	0.00% <0.00%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
packages/loopover-mcp/bin/loopover-mcp.ts	13.23% <100.00%> (+1.63%)	⬆️

... and 644 files with indirect coverage changes

[loopover-orb]

Warning

⏸️ LoopOver review result - manual review recommended

_{Review updated: 2026-07-21 23:40:50 UTC}

3 files · 1 AI reviewer · no blockers · CI green · dirty

⏸️ Suggested Action - Manual Review

Review summary
This PR adds the missing local stdio mirror of loopover_watch_issues (list/watch/unwatch), extracting the shared HTTP dispatch into watchIssuesRequest and reusing it in both the new stdio tool and the existing watch CLI, closing #7763 as described. The refactor is correct: the CLI's three call sites now route through the same helper with identical GET/POST/DELETE semantics and the labels-omission behavior is preserved. Test coverage is solid — in-process InMemoryTransport tests exercise all three actions, both label branches, and both throw guards (missing login, missing repoFullName), plus the CLI call sites via watchCli.

Nits — 6 non-blocking

• packages/loopover-mcp/bin/loopover-mcp.ts:2376 uses `async ({ login, action, repoFullName, labels }: any) =>` for the stdio tool handler instead of a typed shape inferred from watchIssuesShape, matching the existing any-typed pattern in the file but still worth tightening if this file's types are ever cleaned up.
• The external brief's 'leaked secret' flag at test/unit/mcp-cli-watch-issues.test.ts:48 is just the literal string `in-process-token` used as a fixture token for the test API client, not a real credential.
• watchIssuesRequest (loopover-mcp.ts:4407) is untyped (`login: any, action: any, ...`) — a small typed union for `action` (`"watch" | "unwatch" | "list"`) would let TypeScript catch a typo in a call site instead of silently falling through to the `list`/GET branch.
• Consider typing watchIssuesRequest's `action` parameter as the same z.enum values used in watchIssuesShape so a future call-site typo can't silently degrade to the GET/list branch.
• If this file's `any`-typed stdio handler pattern is intentional per existing convention (as it appears to be, matching sibling registerStdioTool blocks), no change needed — otherwise consider a shared inferred type from the schema shapes.
• Possible secret-shaped assignment in the diff (generic_secret_assignment) — Verify the value is not a real credential.

Decision drivers

• ✅ Code review — No blockers (1 reviewer)
• ✅ Gate result — Passing (No configured blocker found.)

Context & advisory signals — never blocks the verdict

Signal	Result	Evidence
Linked issue	✅ Linked	#7763
Related work	✅ No active overlap found	No same-issue or scoped active PR overlap found.
Change scope	✅ 20/20	Low review scope from cached public metadata (1 linked issue).
Validation posture	✅ 25/25	PR body includes validation/test evidence.
Contributor workload	✅ 10/10	Author activity: 112 registered-repo PR(s), 51 merged, 19 issue(s).
Contributor context	✅ Confirmed Gittensor contributor	real-venus; Gittensor profile; 112 PR(s), 19 issue(s).
Improvement	✅ Minor	risk: clean · value: minor

Linked issue satisfaction

Addressed
The PR adds a registerStdioTool("loopover_watch_issues", ...) block mirroring the sibling pattern, uses stdioToolDescription for the description, and refactors the HTTP calls into a shared watchIssuesRequest helper reused by both the CLI and the new stdio tool, satisfying the issue's core requirements.

Review context

• Author: real-venus
• Role context: outside_contributor
• Public audience mode: oss maintainer
• Lane context: Repository is configured for direct PR review.
• Public profile languages: TypeScript, JavaScript, Python, Rust, CSS, MDX, Svelte, Swift
• Official Gittensor activity: 112 PR(s), 19 issue(s).
• PR-specific overlap: none found.

Contributor next steps

• Keep the PR focused and include validation evidence before maintainer review.

Signal definitions

• Related work = same linked issue, overlapping active PRs, or title/path similarity.
• Change scope = cached public metadata such as size labels, draft state, and review-burden hints.
• Validation posture = whether the PR provides enough public validation/test evidence for maintainer review.
• Contributor workload = public contributor activity and cleanup pressure, not a repo-wide quality failure.
• Contributor context = public GitHub/Gittensor identity context; non-Gittensor status is not a blocker.

🧪 Chat with LoopOver

Ask LoopOver a question about this PR directly in a comment — grounded only in the same cached, public-safe facts shown above, never a new claim.

• @loopover ask <question> answers contribution-quality Q&A with source citations and freshness.
• @loopover chat <question> answers in natural prose from cached decision-pack facts via local inference (maintainer/collaborator; read-only).
• A plain-language @loopover mention with a real question is routed to the closest matching read-only command automatically — no exact syntax required.

Full command reference: https://loopover.ai/docs/loopover-commands

🧪 Experimental — new and may change.

_{🟩 Safe / merged · 🟦 Advisory · 🟨 Held for review · 🟥 Blocked / closed}

---

💰 Earn for open-source contributions like this. Gittensor lets GitHub contributors earn for the work they already do — register to start earning →.

Checked by LoopOver, a quiet PR intelligence layer for OSS maintainers.

• Re-run LoopOver review

[loopover-orb]

LoopOver approves — the gate is satisfied and CI is green.

[loopover-orb]

LoopOver is closing this pull request on the maintainer's behalf (conflicts with the base branch — resolve and open a fresh PR). This is an automated maintenance action — to pursue this change, please open a new pull request with the issues resolved. Closed PRs may be analyzed later to improve review accuracy, but they are not automatically reopened or re-reviewed.

[JSONbored]

Merge conflicts, a PR before yours touched the same file:

This branch has conflicts that must be resolved
Use the [web editor](https://github.com/JSONbored/loopover/pull/7959/conflicts) or the command line to resolve conflicts before continuing.

test/unit/mcp-tool-rename-aliases.test.ts