Agent skill
search-mcp-github
Search MCP PRs, issues, and discussions across the modelcontextprotocol GitHub org
Install this agent skill to your Project
npx add-skill https://github.com/modelcontextprotocol/modelcontextprotocol/tree/main/plugins/mcp-spec/skills/search-mcp-github
SKILL.md
Searching MCP PRs, issues, and discussions
Where to search
- MCP Docs Server (
mcp-docsMCP server →SearchModelContextProtocoltool): Authoritative for current spec content. Prefer this first for specification details, API references, and protocol concepts. - Spec PRs & Issues:
gh search prs/gh search issuesinmodelcontextprotocol/modelcontextprotocol(searches open and closed by default) - Spec Discussions: https://github.com/modelcontextprotocol/modelcontextprotocol/discussions (requires GraphQL — see below)
- Org Discussions: https://github.com/orgs/modelcontextprotocol/discussions (requires GraphQL — see below)
For historical decisions, prioritize merged PRs and closed issues over open items.
Searching discussions
There is no gh search discussions command. Use the GraphQL API:
# Spec-repo discussions
gh api graphql -f query="query { search(query: \"repo:modelcontextprotocol/modelcontextprotocol <topic>\", type: DISCUSSION, first: 20) { nodes { ... on Discussion { title url body author { login } authorAssociation category { name } answer { author { login } authorAssociation body } } } } }"
# Org-wide discussions
gh api graphql -f query="query { search(query: \"org:modelcontextprotocol <topic>\", type: DISCUSSION, first: 20) { nodes { ... on Discussion { title url body author { login } authorAssociation category { name } answer { author { login } authorAssociation body } } } } }"
Search term variants
GitHub search does not split camelCase tokens. ToolAnnotations and Tool Annotations return almost entirely different results — search both.
- camelCase (
ToolAnnotations,inputSchema): matches identifiers in code and schema - Space-separated (
Tool Annotations,input schema): matches natural-language discussion text
Skip kebab-case variants (tool-annotations) — GitHub tokenizes on hyphens, so they behave like the space-separated form but tend to return noisier results.
Deep diving into a PR
When to deep dive: a search result PR looks highly relevant to the topic, and you need to understand why a change was made, not just what changed
During a deep dive, look through:
- general conversation on the PR not tied to specific lines of code (
repos/modelcontextprotocol/modelcontextprotocol/issues/{pr_number}/comments) - comments left on specific lines of code during review (
repos/modelcontextprotocol/modelcontextprotocol/pulls/{pr_number}/comments) - top-level review bodies submitted with an approve/request-changes/comment verdict (
repos/modelcontextprotocol/modelcontextprotocol/pulls/{pr_number}/reviews)
Each comment returned by these endpoints includes an author_association field — use it to identify maintainers (see Notable maintainer quotes).
Output format
PRs
- [#123](url) - PR Title (**Merged/Closed/Open** <date>)
Brief summary of PR
Issues
- [#456](url) - Issue Title (**Open/Closed** <date>)
Brief summary of issue
Discussions
- [#789](url) - Discussion Title (<date>)
Brief summary of discussion content
Notable maintainer quotes
Identifying maintainers: The GitHub API includes an author_association field (REST) or authorAssociation (GraphQL) on every comment. Treat users with association MEMBER or OWNER as maintainers.
When maintainers make comments that reveal design intent, set direction, or explain rationale, quote them directly with attribution and a footnote:
"These would require a SEP. I think the general question here is about the taxonomy of hints." [^1] — @dsp-ant
Look for quotes that:
- Explain why a decision was made
- Set direction for future work
- Reject or redirect an approach
- Clarify the intended semantics of a feature
Key insights
Summarize the most important findings and any decisions or consensus reached.
Footnotes
Collect all sources as footnotes at the end. Every quote and claim presented in the output should have a corresponding footnote. For example:
[^1]: [#616 inline review comment by @dsp-ant](https://github.com/modelcontextprotocol/modelcontextprotocol/pull/616#discussion_r...)
[^2]: [#185 ToolAnnotations](https://github.com/modelcontextprotocol/modelcontextprotocol/pull/185)
[^3]: [Spec: Tool Annotations (2025-11-25)](https://modelcontextprotocol.io/specification/2025-11-25/server/tools)
General strategy
- Generate search terms and variants (camelCase, space-separated, etc.)
- Use
SearchModelContextProtocoltool (if available) to search for current specification details and concepts - Expand search terms and variants based on new information
- Search GitHub locations (use
ghCLI tool if available) - Aggregate search results
- Display output with summarized results, key insights, and direct attributions
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
convert-web-app
This skill should be used when the user asks to "add MCP App support to my web app", "turn my web app into a hybrid MCP App", "make my web page work as an MCP App too", "wrap my existing UI as an MCP App", "convert iframe embed to MCP App", "turn my SPA into an MCP App", or needs to add MCP App support to an existing web application while keeping it working standalone. Provides guidance for analyzing existing web apps and creating a hybrid web + MCP App with server-side tool and resource registration.
add-app-to-server
This skill should be used when the user asks to "add an app to my MCP server", "add UI to my MCP server", "add a view to my MCP tool", "enrich MCP tools with UI", "add interactive UI to existing server", "add MCP Apps to my server", or needs to add interactive UI capabilities to an existing MCP server that already has tools. Provides guidance for analyzing existing tools and adding MCP Apps UI resources.
migrate-oai-app
This skill should be used when the user asks to "migrate from OpenAI Apps SDK", "convert OpenAI App to MCP", "port from window.openai", "migrate from skybridge", "convert openai/outputTemplate", or needs guidance on converting OpenAI Apps SDK applications to MCP Apps SDK. Provides step-by-step migration guidance with API mapping tables.
create-mcp-app
This skill should be used when the user asks to "create an MCP App", "add a UI to an MCP tool", "build an interactive MCP View", "scaffold an MCP App", or needs guidance on MCP Apps SDK patterns, UI-resource registration, MCP App lifecycle, or host integration. Provides comprehensive guidance for building MCP Apps with interactive UIs.
kdoc
Add KDoc documentation to Kotlin public API. Use whenever the user asks to document Kotlin code, add KDoc, generate API docs, mentions undocumented public declarations, or wants to improve existing documentation. Also trigger when the user says 'add docs', 'document this class/file/module', 'write KDoc', or asks about missing documentation in Kotlin code.
mcp-docs
Fetch live MCP specification and docs from modelcontextprotocol.io. Training data may be outdated — default to using this skill when the query touches MCP protocol in any way. Trigger for: implementing or debugging MCP features (sampling, elicitation, completion, tools, resources, prompts, transports); SEPs or spec requirements; MCP transport behavior (SSE, Streamable HTTP, reconnection, retry); authorization/OAuth; conformance test failures; capability negotiation; message schemas. Skip only for pure Kotlin, build system, or refactoring tasks with zero MCP protocol dependency. When in doubt, trigger.
Didn't find tool you were looking for?