agentleFS
aradmanamnaoonaradmanamnaoon/vibe-coding-skills/skills/agent-reliability/mcp-reliability-auditor/SKILL.md

mcp-reliability-auditor

Skill0 starsChanged 19 days agoMITView on GitHub

Audit MCP/plugin tool integrations for startup cost, health, retries, timeouts, process leaks, permission scope, and failure isolation. Use when agents stall, spawn processes, or load many MCP servers.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 7 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Keep external tool integrations observable, bounded, and isolated from unrelated work.

    line 8
  2. Distinctive:

    No external tool server is used in the workflow.

    line 16
  3. Distinctive:

    Read applicable repository instructions before edits.

    line 20
  4. Distinctive:

    Start/check servers individually where safe to isolate failures before testing the full client.

    line 35
  5. Distinctive:

    Check cancellation/timeouts and whether a failed server blocks unrelated chat/tool operations.

    line 37
  6. Distinctive:

    Do not expose MCP credentials in reports or disable sandbox/host restrictions merely to make a connection succeed.

    line 47
  7. Distinctive:

    If client internals cannot expose process/queue state, test externally observable startup/timeouts and state the telemetry limitation.

    line 64

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Purpose

Keep external tool integrations observable, bounded, and isolated from unrelated work.

When to activate

An agent uses one or more MCP/plugin servers or tool startup/retry behavior is problematic.

When not to activate

No external tool server is used in the workflow.

Preconditions

  • Read applicable repository instructions before edits.
  • Preserve unrelated working-tree changes and establish a baseline when the workflow can modify code.
  • Identify which tools and environments are actually available; do not claim unavailable checks ran.

Required inputs

  • MCP/client config
  • Server inventory and launch commands
  • Observed process/log/latency data
  • Permission roots/scopes

Workflow

  1. Inventory configured servers and record which are actually needed for the current workflow.
  2. For each needed server, record transport, startup command/URL, permissions, credential source by name only, expected startup timeout, and retry policy.
  3. Start/check servers individually where safe to isolate failures before testing the full client.
  4. Measure/observe startup failure behavior and ensure permanent configuration errors do not trigger unbounded retries/process spawning.
  5. Check cancellation/timeouts and whether a failed server blocks unrelated chat/tool operations.
  6. Reduce filesystem/network/account permissions and disable unused servers for the task when configuration is project-scoped and authorized.
  7. Record versions because protocol/server behavior changes quickly.

Relevant tools

MCP config/server logs, process list, client logs, health/tool listing.

Safety restrictions

Do not expose MCP credentials in reports or disable sandbox/host restrictions merely to make a connection succeed.

Validation

  • Each required server has an observed or explicitly unverified health state.
  • Permanent errors have bounded retries and no runaway descendant-process growth in the test.
  • Unused/high-impact permissions are removed or documented.

Expected output

  • MCP inventory
  • Health/startup evidence
  • Retry/timeout/process findings
  • Permission-scope recommendations

Failure handling

If client internals cannot expose process/queue state, test externally observable startup/timeouts and state the telemetry limitation.

Completion criteria

The skill is complete only when every applicable validation item above has observed evidence or is explicitly marked not executed with the concrete blocker. Do not replace missing evidence with confidence language.

Research mapping

This skill was selected from research problems: P055, P007. See research/vibe-coding-problems.md and research/problem-evidence.csv in the Vibe Coding Skills repository.

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.