docs(npm): fix stale MCP tool count in package README

[nhtkylc91]

Summary

• `pkg/npm/README.md` claimed 14 MCP tools, but the actual `TOOLS[]` registry in `src/mcp/mcp.c` defines 15 (verified by counting entries in the array), matching the count already stated in the root `README.md`.
• Fixes both occurrences in the npm package README so it's consistent with the source of truth and the other docs.

Test plan

• Counted tool definitions in `src/mcp/mcp.c`'s `TOOLS[]` array: 15.
• Confirmed root `README.md` and `pkg/npm/README.md` now agree (15).
• Docs-only change, no code/behavior affected.

[DeusData]

Thank you for the contribution and for checking the package README against the tool registry. This is now triaged as a normal-priority documentation correction for 0.9.1-rc. The tool-count point overlaps the newer #1297 and #1315 report and will need to be reconciled against the actual tools/list response during review. Our community PR queue is currently quite full, so it may take a little time before we can complete that review. We are doing our best to support community contributions and will return with code-grounded feedback as capacity opens.

[DeusData]

Thank you — this one is simply correct. The count is 15: check_index_coverage shipped as the 15th tool in v0.9.1, and the npm README still said 14.

(Amusing companion note: another PR is currently proposing the opposite change to the main README, because it was matching --help — which was itself omitting that same 15th tool. #1370 fixes the help text to derive from the tool registry so the two cannot drift again. Yours is the version that matches reality.)

I cannot merge it yet, and the reason is on our side rather than yours. main requires two status checks, dco and ci-ok, and neither ran on this PR — the only file it touches, pkg/npm/README.md, sits outside the path filters that trigger our workflows, so the required checks never report and GitHub holds the merge as BLOCKED.

If you would not mind: rebase onto current main and push the branch again. That normally re-triggers the workflow set so the checks can report, after which this goes straight in.

If it still shows no checks afterwards, please say so and leave it with me — at that point it is our path-filter configuration that needs fixing rather than anything about your patch, and I will not ask you to keep poking at it. Thanks again for catching this.

[DeusData]

Thank you — you are right, and I verified it independently rather than taking the number on trust.

TOOLS[] in src/mcp/mcp.c defines exactly 15 tools today: index_repository, search_graph, query_graph, trace_path, get_code_snippet, get_graph_schema, get_architecture, search_code, list_projects, delete_project, index_status, check_index_coverage, detect_changes, manage_adr, ingest_traces. I counted it on current main, including after a commit that touched mcp.c more recently than your PR — still 15. The root README.md already says 15, so the npm package README had simply drifted. Your change is correct on both lines.

One thing is needed from you before it can land: a DCO sign-off.

Every commit here needs a Signed-off-by: trailer matching the author, and commit 881eec73 does not have one. That is enforced by CI, so it would fail the moment the checks run. The fix:

git commit --amend -s
git push --force-with-lease

The sign-off email needs to match your commit author email.

And one thing that is on us, not you. The CI checks have not run on this PR at all yet — they are showing action_required, which means they are waiting on a maintainer to approve workflow runs for a first-time contributor. That is a GitHub default, not anything you did wrong, and it is why this PR has looked stuck with no checks rather than with failing ones. We will trigger that once the sign-off is in, so you get a clean run rather than a red DCO followed by a second round.

A small heads-up, explicitly not something to fix here: the same file also says "159 languages" while the root README currently says 158, and our own grammar manifest says 159. That inconsistency predates you and is ours to sort out — please leave it alone in this PR so the diff stays exactly what it claims to be.

Thanks for catching the drift. Small documentation corrections like this are genuinely useful, because the npm README is what most people actually read first.

[DeusData]

Approved the workflow runs — CI is executing now. Apologies that this sat: your checks were never failing, they were waiting on a first-time-contributor approval, which is on us rather than on you.

And you are right on the facts. I counted the registry on current main: TOOLS[] in src/mcp/mcp.c has exactly 15 entries, and the one the README was missing is check_index_coverage. So both edits are correct.

One thing will still be red, and it needs you: the dco check. Your commit has no Signed-off-by: line, and that check is required. For a single commit, git commit --amend --signoff followed by a lease-guarded force push to your branch will do it. The sign-off has to use the same name and email as the commit author, which --signoff handles for you.

Once that lands green it can go in. Thank you for the fix — a stale number in a package README is exactly the kind of thing that quietly erodes trust in the rest of the document, and it is the sort of contribution that often goes unmade because it feels too small to bother with. It isn't.