agentleFS
wnkincwnkinc/claude-custom-connector-server/.claude/skills/verify/SKILL.md

verify

Skill1 starsChanged 4 months agoMITView on GitHub

Run this repo's MCP servers + approval sidecar locally and drive them over real MCP HTTP to verify a change end-to-end.

  • Installs packages
  • Sends data out

189 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use instinct-export, instinct-import and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • TheKREATORSSkill

    instinct-export

    Export instincts from project/global scope to a file

    Also in 3 projects that use this

  • TheKREATORSSkill

    instinct-import

    Import instincts from file or URL into project/global scope

    Also in 3 projects that use this

  • AviadPSkill

    risk-assess

    Assess the risk and blast radius of a proposed code change before implementing it. Evaluates CI impact, framework-wide vs localized scope, rollback path, and identifies blindspots. Use when user says "what's the risk", "is this safe", "assess the impact", "risk assess", "blast radius", or before applying framework-wide or CI-affecting changes. Do NOT use for code quality review (use code-quality) or plan validation (use design-validator).

    Also in 3 projects that use this

  • diego-tobalinaSkill

    build-fix

    Fix build and compilation errors systematically. Run when build fails.

    Also in 6 projects that use this

  • AaravChadhaSkill

    secure

    Security review with a confidence gate - a finding exists only with a concrete exploit scenario and a high/medium/low confidence rating; everything else goes to a worth-hardening list, never inflated. Sweeps five surfaces: auth gates, secrets hygiene, injection and unsafe sinks, unsafe deserialization/crypto/transport, and LLM tool-use trust boundaries. Reports only, never fixes. Use when the user asks for a security review or to check vulnerabilities, secrets, or auth.

    Also in 3 projects that use this

  • ArvorCoSkill

    specify

    Create feature specification from natural language description. Use when starting a new feature or creating a spec. Triggers on: create spec, specify feature, new feature spec.

    Also in 10 projects that use this

  • +1 more from TheKREATORS

The file

Files
SKILL.md

Verify an mcp-tools change at the MCP surface

The surface is MCP over HTTP: a tool server composed by security/serve.py, talking to the approval sidecar. Drive it with a fastmcp.Client, not unit tests.

Environment

uv venv /tmp/verify-env && VIRTUAL_ENV=/tmp/verify-env uv pip install "fastmcp==3.4.2" uvicorn pynacl
PY=/tmp/verify-env/bin/python

(fastmcp==3.4.2 matches the tools' locks; pynacl is for the sidecar's Discord path import.)

Launch (from the repo root; pick free ports)

APPROVAL_PUBLIC_URL=http://127.0.0.1:18072 APPROVAL_PORT=18072 $PY security/approval/service/service.py &
APPROVAL_URL=http://127.0.0.1:18072 APPROVAL_PUBLIC_URL=http://127.0.0.1:18072 \
  APPROVAL_WIDGET=1 MCP_PORT=18065 $PY tools/gatekeeper/server.py &
curl -s http://127.0.0.1:18072/healthz   # sidecar up; provider slack/unconfigured is fine locally
  • Auth is off by default locally (MCP_AUTH_ENABLED unset); no OAuth needed.
  • The real telegram server needs live Telegram credentials + the vendored stdio engine. For changes in the shared layers (serve/middleware/gating), stand in a minimal FastMCP(name="telegram") with one write + one read tool and call serve(mcp, port=..., require_approval=True) — same composition path, same source name. Tools ship always_allow; to exercise the gate, set a mode first: curl -XPOST {sidecar}/gating -d '{"source":"telegram","tool":"<write tool>","mode":"needs_approval"}' -H 'content-type: application/json' (modes cache for ~15s in the server).

Drive

fastmcp.Client("http://127.0.0.1:<port>/mcp") → list_tools() / call_tool().

  • Gated call round-trip without a chat channel: run servers with APPROVAL_WIDGET=1; the pending text carries <!--APPROVAL {"token": ...}--> — parse the token and POST {sidecar}/approve/{token} with form decision=approve (what the human's browser does), then re-call the tool with the same args.
  • Gating/mode overrides: tool servers cache the sidecar's /gating for 15s — sleep 16s after flipping a mode before asserting list/call behavior.

Gotchas

  • A gated tool's FIRST call short-circuits before FastMCP argument validation, so bad args still mint an approval card; validation errors only surface on the post-approval re-call.
  • Approvals are one-shot: each approved call consumes its record; the next identical call mints a fresh card.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.