Generic markdown vault MCP with hybrid search Skills live under .agents/skills/; Claude Code reads them through the .claude/skills/ symlinks. They load only when invoked, so invoke the one that matches before you start:
Use when filing a bug, opening or creating a GitHub issue, drafting an epic or ticket, writing up a finding or review observation worth tracking, sizing a change before coding, or opening a pull request for this repository. Routes the change to the right repo first (library / template / domain), picks the right issue form, sizes the change to one PR, links epic children as native sub-issues, and applies CONTRIBUTING.md before anything is posted.
Use when writing or changing an MCP tool that can fail, refuse, find nothing or hit a conflict: choosing between returning a result and raising ToolError, writing a tool's error text, choosing log levels for tool outcomes, or when a tool call logs ERROR or a traceback for something that is not a server fault.
Use before any release work: the trunk release model, the unstable edge channel, the prepare/release/promotion machinery, release-notes pages, the pre-release artifact smoke test, and, when the plugin channel is enabled, the Claude Code plugin channel.
Use when a change depends on how something outside this repository actually behaves — a markdown dialect, a file format, git, a protocol, a vendor API — and docs/design/reference/ has no current reference for it, or the one it has is past its review date. Researches from primary sources and writes a dated, versioned, source-marked reference so the next agent reads instead of re-deriving. Works with any coding agent; parallel research is optional.
Use when work spans more than one pull request or epic, when the order of work is not yet decided, or when charting, refining or revisiting the roadmap index, epics and release package milestones. Use before implementation planning for work not yet refined into one ready feature.
Use before opening a pull request, marking one ready for review, or pushing further commits to a branch with an open pull request — or when asked to review a diff or a PR. Reviews the cumulative diff against the project's written commitments and reports evidence-grounded findings; works with any coding agent and needs no hosted model workflow.
Use when adding, renaming, or documenting an MCP tool: the registration checklist, the get_server_info tool, tool icons, and the public import-surface guard.
Use before adding, moving or substantially changing a page in docs/ or README.md, or a DOMAIN block in a template docs page: decides whether the knowledge belongs to this project or to the template, which existing page or designated place takes it, and what the change must link to.
Use when writing or changing a tool, parameter, resource or prompt description, a tool title or annotation, or a server-instructions snippet, and when reviewing such text: it decides which sentences the model can act on and where every other fact goes instead.
Project instructions live in AGENTS.md (domain content between its DOMAIN-START / DOMAIN-END markers). This file is template-owned; do not add content here.
Scholarly papers, patents, books, standards and PDF conversion Skills live under .agents/skills/; Claude Code reads them through the .claude/skills/ symlinks. They load only when invoked, so invoke the one that matches before you start:
Use before any release work: the trunk release model, the unstable edge channel, the prepare/release/promotion machinery, release-notes pages, the pre-release artifact smoke test, and, when the plugin channel is enabled, the Claude Code plugin channel.
Use when adding, renaming, or documenting an MCP tool: the registration checklist, the get_server_info tool, tool icons, and the public import-surface guard.
Paperless-NGX over MCP: search, read, upload and tag documents; manage correspondents and types. Skills live under .agents/skills/; Claude Code reads them through the .claude/skills/ symlinks. They load only when invoked, so invoke the one that matches before you start: Project-owned skills follow the same shape: a directory under .agents/skills/ plus a relative symlink in .claude/skills/. Project-specific issue and PR conventions go inside the authoring-issues-prs skill's DOMAIN-AUTHORING block. Mark a commit ! (or add a `BREAKI
Use before any release work: the trunk release model, the unstable edge channel, the prepare/release/promotion machinery, release-notes pages, the pre-release artifact smoke test, and, when the plugin channel is enabled, the Claude Code plugin channel.
Use when adding, renaming, or documenting an MCP tool: the registration checklist, the get_server_info tool, tool icons, and the public import-surface guard.
Plain text files in a repository that tell a coding agent how the project works: commands to run, conventions to follow and things to avoid. CLAUDE.md, AGENTS.md, cursor rules and skills are the common kinds.
CLAUDE.md or AGENTS.md?
CLAUDE.md is read by Claude Code. AGENTS.md is an open format that Codex, Cursor and other agents read. Many projects keep one and point the other at it.
What is a skill?
A folder with a SKILL.md that describes one capability, such as filling PDFs or reviewing code. The agent loads it only when the task calls for it.
Can I search my own team's files too?
Your agents already can, over MCP, limited to the files you're allowed to read. Searching them from this page is coming.