Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@skill-lint lint my skills folder and show warnings"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Lint agent SKILL.md files and MCP tool schemas.
Agent skills and MCP tools rarely fail loudly. They fail by never being chosen —
a description that never says when to use it, two tools the model can't tell
apart, a schema so verbose it costs you tokens on every single turn. skill-lint
finds those, plus the ones that bite harder: a destructive tool with no
confirmation gate, an unconstrained command parameter, a skill pointing at a
file that isn't there.
Zero dependencies. Runs as a CLI or as an MCP server.
uv tool install skill-lint # or: pipx install skill-lint
skill-lint ~/.claude/skills # lint a skill library
skill-lint --mcp tools.json # lint an MCP tools/list payload
skill-lint . --format github # CI annotations$ skill-lint ~/.claude/skills
skills/report-writer/SKILL.md
▲ SL007 description never says WHEN to use the skill — Generates polished reports…
→ Add an explicit trigger clause: "Use when …", "Triggers on …". This is the
single most common reason a working skill never gets invoked.
▲ SL011 references a missing file — templates/quarterly.md
→ The model will try to read this and fail.
197 checked · 1 error(s) · 129 warning(s) · 75 infoRelated MCP server: policy-gate-mcp
// claude_desktop_config.json / .mcp.json
{
"mcpServers": {
"skill-lint": { "command": "skill-lint-mcp" }
}
}Exposes lint_skills and lint_mcp_tools. The server speaks the MCP stdio
protocol directly — no SDK, so it starts in milliseconds and adds nothing to
your dependency tree.
Code | Severity | What it catches |
SL001 / SL002 | error | Missing |
SL003 | warn |
|
SL004 | warn | Name isn't kebab-case, or is over 64 chars |
SL005 / SL006 | warn | Description too short to route on / long enough to cost real tokens |
SL007 | warn | Description never says when to use the skill |
SL008 | info | Description written in first person |
SL009 | warn/error | Skill body large enough to bloat every load |
SL010 | error | Frontmatter with no body |
SL011 | warn | Links to or runs a file that doesn't exist |
SL012 | error | Two skills share a name — one silently shadows the other |
SL013 | error | Frontmatter doesn't parse |
SL014 | info | Unrecognised frontmatter keys (usually typos) |
SL016 | warn | Two descriptions so similar the model can't choose between them |
SL007 is the one that matters most. On a real 197-skill library it fired 116 times — most skills describe what they are and never say when to reach for them, so the router never picks them and the author assumes the skill is fine.
Trigger detection is English-first. Descriptions that don't look English are
reported at info rather than warn, since the heuristic can't judge them.
Code | Severity | What it catches |
MC001 | error/warn | Tool has no description, or one too short to route on |
MC002 | warn | Parameter has no description |
MC003 | error | Destructive tool with no confirmation gate |
MC004 | warn | Unconstrained free-form |
MC005 | info | Tool name isn't snake_case |
MC006 | warn | Description long enough to matter, paid every request |
MC007 | warn | Schema has no |
MC008 | info |
|
MC009 | error | Duplicate tool name — one is unreachable |
MC010 | warn | More than ~40 tools; selection accuracy drops |
MC011 | warn | Whole tool list is expensive per request |
MC003 matches the tool name and its opening sentence, not the whole description — a destructive verb in later context ("use before publishing") isn't what the tool does, and matching it turns the rule into noise.
--mcp treat paths as MCP tool-schema JSON
--format text|json|github
--fail-on error|warn|info|never minimum severity that exits non-zero
--select SL007,MC003 only these codes
--ignore SL014 suppress these codes- run: pipx install skill-lint
- run: skill-lint .claude/skills --format github --fail-on warnPYTHONPATH=src python3 -m unittest discover -s tests -vThe test suite includes a dogfood case: this server's own tool schema has to pass its own linter.
MIT
Lint an MCP tool schema — pass the JSON from a tools/list response or a bare array of tool objects. Reports missing descriptions, destructive tools with no confirmation gate, unconstrained free-form parameters that widen prompt-injection surface, duplicate tool names, and the per-request token cost of the whole tool list. Use when designing or reviewing an MCP server.
| Name | Required | Description | Default |
|---|---|---|---|
| schema | Yes | The tools/list JSON payload, as a string. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries the full burden and provides a rich list of what it reports: missing descriptions, destructive tools without confirmation gates, unconstrained parameters, duplicate names, and token cost. It does not explicitly state that the operation is read-only or describe the report format, but the behavioral scope is well disclosed.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Three sentences, front-loaded with purpose, then output details, then usage context. No filler or repetition.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a single-parameter lint tool with no output schema or annotations, the description adequately explains what the tool inspects and when to use it. It does not cover error behavior or output structure in detail, but the report categories give sufficient completeness.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100% and the single parameter is documented. The description adds meaning beyond the schema by clarifying that the input can be a tools/list response or a bare array of tool objects, which is useful format guidance.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
States a specific verb+resource ('Lint an MCP tool schema') and clearly identifies the input sources. It does not explicitly distinguish itself from the sibling lint_skills tool, though the resource domain is distinct.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Provides a clear usage context: 'Use when designing or reviewing an MCP server.' It does not state when not to use it or name alternative tools.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Lint agent SKILL.md files in one or more directories. Reports missing or untriggerable descriptions, name/directory mismatches, oversized bodies, broken file references, duplicate skill names, and pairs of skills too similar for a model to choose between. Use when auditing a skill library, before publishing a skill, or when a skill exists but never seems to get invoked.
| Name | Required | Description | Default |
|---|---|---|---|
| paths | Yes | Directories to search recursively for SKILL.md files. | |
| ignore | No | Rule codes to suppress, e.g. SL014. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries the full behavioral burden. It helpfully discloses the categories of findings (untriggerable descriptions, name mismatches, oversized bodies, broken references, duplicates, near-duplicates), which tells the agent what a run surfaces. It never states that the operation is read-only/non-mutating, nor describes output shape or severity, so key traits remain inferred.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Three sentences, tightly front-loaded: what it lints, what it reports, when to use it. The detection list is long but every item is a distinct, agent-relevant finding category, so nothing is padding.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a two-parameter diagnostic tool with no output schema, the description is nearly sufficient: scope, detection semantics, and trigger conditions are all present. It stops short of saying whether files are modified, how findings are returned, or how to act on them.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, so both parameters are already documented, and the baseline is 3. The description only echoes the directory scope ('one or more directories'); the 'ignore' rule-code parameter and the SL### format are never mentioned in prose.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
Specific verb ('Lint') and resource ('agent SKILL.md files in one or more directories'), plus an enumeration of exactly what problems it detects. It is clearly distinguishable from the sibling lint_mcp_tools by resource (SKILL.md vs MCP tool definitions), but it never names or contrasts with that sibling explicitly.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Gives three concrete triggering situations: auditing a library, pre-publish checks, and diagnosing skills that never get invoked. That is genuine when-to-use guidance. It does not state when NOT to use it or point to lint_mcp_tools for non-skill linting.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
v0.1.0lint_mcp_toolslint_skillsScored across 2 tools
The two tools are cleanly separated by input target: lint_skills operates on SKILL.md files/directories, while lint_mcp_tools operates on MCP tool schema JSON. There is no realistic way to confuse the two, since each description names a distinct artifact and use case.
Both names follow an identical verb_noun pattern (lint_skills, lint_mcp_tools) with consistent snake_case and a shared 'lint_' prefix that signals the server's purpose. Fully predictable.
Two tools map directly to the server's two linting domains, so each earns its place and the surface is tightly scoped. It is slightly thin—there is no config, list, or fix operation—but not unreasonably so for a focused linter.
The linter covers both advertised domains and reports a broad range of issues (descriptions, naming, size, duplicates, similarity, token cost). Missing auto-fix or rule-configuration operations are minor gaps an agent can work around.
Lint a SKILL.md for frontmatter, structure, secrets and size. All 6 tools free.
Statically audits MCP tool surfaces for token cost, schema quality, and design issues.
MEOK AGENTS.md Linter MCP — validates the cross-vendor coding-agent spec (Cursor / Claude Code /
Search, fetch, lint, and install Agent Skills (SKILL.md) from the SkillMD registry.