skills-doctor
Optionally audits the global skill roots installed for Hermes (skipped by default, enabled with --include-other-agents), including it in the shadowing, broken-reference and token-budget analysis.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
openclaw-skills-doctor
Audit the skills OpenClaw loads: the per-step token tax, shadowed duplicates, broken references and budget overruns - then get a measured plan to fix them.
OpenClaw injects every eligible skill into the system prompt as a compact XML block: roughly 97 characters plus name, description and location per skill, paid on every model step. This tool measures that cost, tells you which definition actually wins, and can trim or hide the expensive ones.
A thin, focused wrapper around opencode-skills-doctor
that knows OpenClaw's skill roots, precedence, openclaw.json, its MCP registry
and the bundled skill.
Quick start
npx -y openclaw-skills-doctor check # findings + per-step token cost
npx -y openclaw-skills-doctor plan # exact savings: trim and/or hide
npx -y openclaw-skills-doctor doctor # paths, roots, install state
npx -y openclaw-skills-doctor check --html openclaw-report.html --openNothing to configure: the wrapper resolves the state dir and workspace the same way OpenClaw does.
Related MCP server: whats-loaded-mcp
What it audits
OpenClaw file-backed skill roots, highest precedence first:
Priority | Source | Path |
1 | Workspace skills |
|
2 | Project agent skills |
|
3 | Personal agent skills |
|
4 | Managed / local skills |
|
5+ | Extra dirs |
|
The wrapper hands those roots to the core auditor in exactly this order, so the
winning definition in a shadowing report matches what OpenClaw actually loads.
Other agents' global roots (Claude, OpenCode, Grok, Cursor, Kimi, Hermes) are
skipped by default; pass --include-other-agents to audit them too.
Install the integration
npx -y openclaw-skills-doctor installThat does two things and is idempotent:
saves the MCP server definition
mcp.servers.skills-doctorin~/.openclaw/openclaw.json- throughopenclaw mcp addwhen the CLI is installed, otherwise by patching the config directly (the previous file is backed up asopenclaw.json.bak-<timestamp>);installs the bundled
openclaw-skills-doctorskill into<workspace>/skills/openclaw-skills-doctor.
npx -y openclaw-skills-doctor install --dry-run # show what would change
npx -y openclaw-skills-doctor uninstall # remove both again
openclaw mcp doctor skills-doctor --probe # verify the server connectsRestart the Gateway (or run openclaw mcp reload) so the tool appears. MCP tools
are namespaced per server, so the agent calls skills-doctor__skills_doctor.
Manual MCP setup
If you prefer to wire it yourself:
openclaw mcp add skills-doctor \
--command npx \
--arg -y \
--arg opencode-skills-doctor \
--arg mcpor in ~/.openclaw/openclaw.json:
{
mcp: {
servers: {
"skills-doctor": {
command: "npx",
args: ["-y", "opencode-skills-doctor", "mcp"],
},
},
},
}Paths it resolves
What | Resolution order |
State dir |
|
Config |
|
Workspace |
|
Override any of them with --state-dir <dir>, --workspace <dir>, --home <dir>.
Commands
Command | What it does |
| audit sources, shadowing, token budget and broken references (default) |
| exact savings: which descriptions to trim, which skills to hide |
| every skill with its source, precedence and cost |
| archive shadowed duplicates (reversible) or hide low-frequency skills |
| MCP server entry + bundled skill, idempotent |
| CLI, state dir, config, workspace, roots and install state |
| run the MCP server on stdio |
Core flags work as usual: --json, --budget, --lang en|zh, --dry-run,
--strict, --html [file], --no-global.
Fixing what it finds
npx -y openclaw-skills-doctor plan # measure first
npx -y openclaw-skills-doctor fix --dry-run # see what would be archived
npx -y openclaw-skills-doctor fix # archive shadowed duplicates (reversible)
npx -y openclaw-skills-doctor fix --hide legacy-skillfix only archives definitions that never load, moving them to
.skills-doctor-archive/<timestamp>/ inside the same skill root. --hide adds
metadata.opencode/autoinvoke: false, which keeps the skill loadable through an
explicit @id reference while removing it from the model-facing list.
Safety
check,plan,scananddoctorare read-only and offline (no network, no LLM).install/uninstallback upopenclaw.jsonbefore rewriting it; the CLI path leaves the file alone entirely.uninstallonly removes a bundled skill whoseSKILL.mdstill declaresname: openclaw-skills-doctor.Config parsing understands JSON with comments and trailing commas. A file that uses other JSON5-only syntax (unquoted keys, single quotes) is reported, never rewritten - use
openclaw mcp add/set/unsetthere instead.It is not a sandbox: it audits configuration and files, it does not constrain what the agent can do.
Development
npm install
npm test # 19 testsRequires Node >= 20. MIT.
This server cannot be deployed
Maintenance
Related MCP Connectors
Statically audits MCP tool surfaces for token cost, schema quality, and design issues.
Audit MCP configs and skill files for over-broad scope and injection risk. 6 of 8 free.
Vet an agent skill before you install it: structure, triggers, secrets.
Audits any remote agent or MCP URL for JSON-RPC compliance, latency SLA, OpenAPI schema, and HTTP.
Related MCP Servers
- AlicenseNot gradedqualityCmaintenanceAudits AI agent skills for safety using static, semantic, adversarial, and supply-chain analysis, providing scores and risk flags. Can be run via CLI, CI, or as an MCP tool from Claude Code, Cursor, and Codex.2Apache 2.0
- AlicenseAqualityBmaintenanceShows what is consuming your agent context window before a session starts: always-loaded skill descriptions ranked by token cost, skills installed more than once under different names, CLAUDE.md/AGENTS.md sizes including what their @import lines pull in, and MCP servers configured across clients. Read-only: no child processes, no network.546 npmMIT
- AlicenseNot gradedqualityBmaintenanceEnables you to audit your AI agent skills by running each one against an agent that cannot see it, diffing the resulting artifacts, and grading whether each skill genuinely improves, changes nothing, or worsens the output.15 npmMIT
- AlicenseNot gradedqualityCmaintenanceEnables authorized security auditing of AI-agent supply chains and agent-facing surfaces: deterministic local skill-bundle audits against eight attack patterns, secrets scanning, and scope-gated read-only recon of agent endpoints and MCP surfaces.MIT