Readme Doc — Documentation skill for Claude Code
Shared knowledge space for coding agents: a local .readme/ of verified notes that Claude Code and Codex both read.
How to install Readme Doc
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open xhuang9/readme-doc and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Readme Doc does
Shared knowledge space for coding agents: a local .readme/ of verified notes that Claude Code and Codex both read. For small to medium-sized projects.
Alternatives in Documentation
- Project Architect — Documentation-first project planning agent skill 250 ★
- Create Release — Create a tagged GitHub release with structured release notes 238 ★
- Doc Fix — Document a recently solved problem for the knowledge base 209 ★
README
readme-doc
Shared context for coding agents. `readme-doc` builds a local knowledge space inside your project, a `.readme/` folder of small, verified notes, so every agent session starts from what earlier sessions learned instead of re-reading the codebase.
It works with **Claude Code** and **Codex**. Both read the same files, so knowledge one agent writes, the other uses.
Who it's for
Best suited to **small and medium-sized projects**: past the point where an agent can hold the whole codebase in context, and small enough that one person or a small team keeps the notes current. Typical signs you're there:
- Agents keep re-discovering the same gotchas.
- The same question about how a part of the system works comes up every week.
- Your
CLAUDE.mdorAGENTS.mdis turning into a wall of text.
A one-file script rarely needs it. A very large codebase usually needs dedicated documentation tooling.
How the knowledge space works
Three tiers, so an agent reads only what the task needs:
CLAUDE.md "Documentation Sections" table ← Tier 0: the index every agent loads first
└─ .readme/sections/.index.md ← Tier 1: one page per area, listing its notes
└─ .readme/chunks/..md ← Tier 2: the knowledge itself
.readme/docs-manifest.json ← generated index of every section and chunk
- Chunks are small and specific. Each one names real files, a concrete pattern and at least one gotcha from your project.
- Chunks know what they describe. YAML frontmatter records
source_pathsandlast_verified_at. - Stale notes get caught.
doc:batch-expirechecks each chunk against the files it describes and marks outdated ones*.expired.md.doc:updatethen refreshes only those. - One entry file. The index lives in
CLAUDE.md.AGENTS.md, which Codex reads first, is a one-line pointer containingCLAUDE.md, so both agents land on the same table.
Skills
| Skill | What it does | | ---
Related Skills
StickyMD
Native sticky notes for GNOME X11, kept behind regular windows and stored as plain Markdown that Codex, Claude
Petit Poucet
Curated, shared Markdown memory for coding agents (Claude Code, GitHub Copilot CLI): one small Rust MCP server
Agent Doc Skills
Navigable project documentation for Claude Code, Cursor, and Codex — a knowledge base your coding agent can pi
Observational Memory
Local-first shared memory for Claude Code, Codex, OpenCode, Kimi, Grok, Cowork, and Hermes, with bounded obser
ThinkWiki
Install once, talk to your agent. ThinkWiki turns documents and notes into a local Markdown wiki with inbox re
Agent Checkpoint
Repo-local checkpoint and resume skills for coding agents. Handoffs are plain markdown under .agents/checkpoin
Related Agents
Exec Mechanical
SMALL tier at medium effort. Execution agent for mechanical, well-specified edits in an isolated git worktree
Kimi Worker
Forward a coding, diagnosis, or review task to Kimi K3 (Moonshot) through the local kimi-worker.sh runtime. Us
Change Scoper
Break work into small, trunk-sized tasks with binary DoD and safe rollback.