Allows public GitHub repositories to be used as sources of reusable agent skills, providing tools to list available skills and read their contents and resources.
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., "@skills-mcp Find a skill for generating pull request summaries"
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.
A lightweight MCP server for exposing reusable agent skills stored in public GitHub repositories.
list_skills
read_skill
list_skill_resources
read_skill_resource
MCP transport is stdio only.
Related MCP server: Personal Code Skills MCP Server
npx @myriadcodelabs/skills-mcpThe process also starts a loopback-only management UI at:
http://127.0.0.1:3218/The UI is only for local repository management; it is not an HTTP MCP transport. It is a small React interface built with standard shadcn/ui components.
To add a source, paste only the public GitHub repository URL, for example:
https://github.com/muhammadismailkhan0009/dev-skillsThe server derives its internal source identity automatically and uses the repository's default branch.
Skill names do not need to be globally unique. Every discovered skill also exposes a derived scope from its parent repository path. For example, two skills named feature-planning can remain distinct as backend/java and frontend/nextjs, while the canonical skill ID remains the exact machine identity.
Repository configuration is persisted at:
~/.skills-mcp/sources.jsonOptional environment variables:
SKILLS_MCP_DATA_DIR — override the data directory
SKILLS_MCP_UI_PORT — override the management UI port; use 0 for an ephemeral port
V1 supports public GitHub repositories only.
Example:
{
"mcpServers": {
"skills": {
"command": "npx",
"args": ["-y", "@myriadcodelabs/skills-mcp"]
}
}
}npm install
npm run typecheck
npm test
npm run buildThe test suite follows behavior-driven scenarios for repository discovery, source lifecycle, MCP tool behavior, UI behavior, and the stdio CLI entrypoint.
See SPECIFICATIONS.md for the V1 scope.
Source repository:
https://github.com/muhammadismailkhan0009/skills-mcpnpm package:
@myriadcodelabs/skills-mcpThe repository includes:
.github/workflows/ci.yml — verifies pushes and pull requests
.github/workflows/release.yml — creates GitHub Releases from v* tags and publishes missing npm versions
npm trusted-publishing support through GitHub Actions OIDC
Create an empty public repository named skills-mcp under:
muhammadismailkhan0009The local repository already has this remote configured:
git remote -vThen push:
git push -u origin mainThe npm package is scoped as:
@myriadcodelabs/skills-mcpYou must own the @myriadcodelabs npm scope, either as your npm username or as an npm organization where you have publish permission. If you do not own that scope, change the package name before the first publish.
Authenticate locally:
npm login
npm whoamiThe first package version must be published manually because npm trusted publishing can only be configured after the package exists:
npm run release:check
npm publish --access publicAfter the first npm publish, use either the npm website or the npm CLI to trust the GitHub Actions release workflow.
Using npm CLI requires npm 11.15.0 or newer:
npm install --global "npm@^11.15.0"
npm trust github @myriadcodelabs/skills-mcp \
--repo muhammadismailkhan0009/skills-mcp \
--file release.yml \
--allow-publishThe equivalent npm website configuration is:
Provider: GitHub Actions
GitHub user/organization: muhammadismailkhan0009
Repository: skills-mcp
Workflow filename: release.yml
Allow direct publish: enabled
No long-lived NPM_TOKEN GitHub secret is required.
After the first npm version has been published and trusted publishing is configured:
git tag v0.1.0
git push origin v0.1.0The release workflow will:
install dependencies
run typecheck, tests, build, and npm package verification
verify that tag v0.1.0 matches package version 0.1.0
detect that @myriadcodelabs/skills-mcp@0.1.0 already exists on npm and skip duplicate publication
create the GitHub Release with generated release notes
Choose the appropriate semantic-version bump without creating the tag yet:
npm version patch --no-git-tag-version
# or:
npm version minor --no-git-tag-version
# or:
npm version major --no-git-tag-versionRun the full release checks before committing or tagging:
npm run release:checkThen commit the version bump and create the matching tag:
VERSION="$(node -p "require('./package.json').version")"
git add package.json package-lock.json
git commit -m "chore: release v$VERSION"
git tag "v$VERSION"
git push origin main --follow-tagsFor subsequent versions, the GitHub Actions release workflow creates the GitHub Release and publishes the new npm version automatically using OIDC trusted publishing.
List resource files such as references, scripts, and assets for a skill.
| Name | Required | Description | Default |
|---|---|---|---|
| id | Yes | Canonical skill ID |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries full responsibility. It states the tool lists resource files, which implies a read operation, but it does not disclose behavior such as pagination, sorting, error handling for invalid skill IDs, or what the response structure looks like. It only says 'list', which is a minimal behavioral statement.
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?
The description is a single, front-loaded sentence that immediately states the action and subject. It is concise with zero wasted words, listing examples of resource types for clarity.
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 simple list tool with one parameter and no output schema, the description is somewhat minimal. It does not mention the return format (e.g., array of file paths/names) or any edge cases (e.g., what happens if the skill has no resources or the ID is invalid). Given no annotations and no output schema, more detail would improve completeness, but it is not critically inadequate for a straightforward listing operation.
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?
The schema provides 100% description coverage for the 'id' parameter ('Canonical skill ID'). The tool description adds no additional meaning to this parameter beyond what the schema already states. The baseline of 3 is appropriate because the schema fully documents the parameter.
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?
The description clearly states the verb 'List' and the resource 'resource files such as references, scripts, and assets' for a skill. It distinguishes itself from siblings like list_skills (which lists skills themselves) and read_skill_resource (which reads a single resource). The scope is explicit and unambiguous.
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?
The description implies usage: call this when you need to enumerate the resources of a specific skill. However, it does not explicitly state when to prefer this over list_skills or read_skill_resource, nor does it provide any exclusions or prerequisites. No explicit routing to alternatives is given.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
List all available skills and their canonical IDs.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries the burden. 'List' implies a read-only operation, but no explicit statement about side effects or performance characteristics (e.g., pagination, ordering) is given. For a simple enumeration, this is sufficient, but it doesn't add any context beyond the obvious.
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?
A single, tightly worded sentence that front-loads the verb and resource, then specifies the output (canonical IDs). No filler or redundancy.
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 parameterless listing tool with no output schema, the description is complete: it states the action, scope (all skills), and the nature of the result (canonical IDs). An agent has enough to call it correctly without further clarification.
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?
There are zero parameters, and the schema already fully documents this (coverage 100%). The description adds no parameter-specific meaning because there are none. Per the rubric, 0 params yields a baseline of 4, and nothing reduces that.
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?
The description clearly states the action (list), resource (skills), and what is returned (canonical IDs). It distinguishes itself from siblings like read_skill (which targets a specific skill) and list_skill_resources (which presumably lists resources of a skill) by focusing on the top-level enumeration of skills.
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?
The description implies this is the entry point for discovering skills before using read_skill, but it doesn't explicitly state when to use it vs. alternatives. There's no exclusion or alternative naming, though the sibling set makes the role obvious. It's adequate but relies on inference.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Read the complete SKILL.md content for a skill ID.
| Name | Required | Description | Default |
|---|---|---|---|
| id | Yes | Canonical skill ID |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden. It only restates the action ('Read the complete SKILL.md content') without disclosing return format, potential errors, or any side effects. For a read operation this is less risky, but the description adds minimal behavioral context beyond the verb.
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?
A single sentence that is front-loaded with the key action and scope, with no redundant words. Perfectly concise for the tool's simplicity.
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 simple one-parameter read tool without an output schema, the description is mostly sufficient. However, it lacks explicit differentiation from read_skill_resource and does not mention edge cases (e.g., invalid ID) or that the response is the raw markdown content. Given the lack of annotations, a bit more context would improve 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% for the single parameter 'id', so the schema already documents it. The description adds no extra semantics about the parameter, matching the baseline for high schema coverage.
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 (Read) and resource (complete SKILL.md content) with a clear parameter (skill ID). It naturally distinguishes from siblings list_skills, list_skill_resources, and read_skill_resource, which target different objects.
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?
The description implies its use case (reading the main skill content) but does not explicitly contrast it with alternatives like read_skill_resource or list_skills. No when-to-use or when-not-to-use guidance is given, so the agent must infer from the tool name and sibling names.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Read one resource file belonging to a skill.
| Name | Required | Description | Default |
|---|---|---|---|
| id | Yes | Canonical skill ID | |
| path | Yes | Resource path relative to skill root |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description must disclose behavioral traits. It only says 'read,' which implies a non-destructive operation, but it does not mention what the tool returns (e.g., file content), error handling, permissions, or any side effects. This is a significant gap for a tool with no annotation safety profile.
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?
The description is a single, front-loaded sentence with zero waste. It efficiently states the tool's purpose without redundant elaboration.
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?
While the tool is simple with two fully documented parameters, the absence of an output schema and any description of the return format or potential errors leaves an agent uncertain about the result. The description is too minimal to be fully actionable on its own.
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 'id' and 'path' are already documented in the input schema. The description adds no additional meaning about the parameters, but since the schema fully covers them, the baseline of 3 is appropriate.
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?
The description states a clear verb and resource: 'Read one resource file belonging to a skill.' It unambiguously identifies the operation and distinguishes it from listing skills or resources. However, it does not explicitly name sibling tools or contrast them, so it stops short of a 5.
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?
There is no guidance on when to use this tool versus alternatives like list_skill_resources or read_skill. The description implies usage through the name and parameters, but no explicit context or exclusions are provided.
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.
v0.1.1list_skill_resourceslist_skillsread_skillread_skill_resourceScored across 4 tools
Each tool targets a distinct action and resource type: listing vs reading, and skills vs resources. No overlap or ambiguity between the four tools.
All tool names follow a consistent verb_noun pattern (list_skills, read_skill, list_skill_resources, read_skill_resource), with clear and predictable conventions.
Four tools perfectly cover the read-only skills management scope without redundancy or excessive granularity, making the set well-scoped.
The tool surface fully covers the apparent domain: listing and reading skills and their resources. For a read-only server, there are no gaps or missing operations.
Search your team's shared AI-skill library, get install commands, and save skills from your agent.
Search and fetch AI agent skills, rules files and MCP servers indexed from GitHub.
Find, inspect and install reviewed reusable AI-agent capabilities from superskill.sh.
Search verified Claude Code plugins and skills; fetch portable SKILL.md sources. Read-only.
