Project Butler
Cross-session project memory for AI coding agents: session logs, handoff, TODO, wiki, status
Test report
- Verdict
- Tested · Works
- Score
- Tested
- Jul 31, 2026
- Environment
- Claude Code 2.x (agent harness)
- Upstream re-checked
- Aug 10, 2026 · 4969a83
Fetched root SKILL.md; spot-checked references/update-check.md, scripts/check-update.sh, references/file-templates.md, references/project-profile-system.md (all HTTP 200). Read check-update.sh fully: it only does git fetch/pull on its own dir and greps CLAUDE.md for a language token — no curl|sh, base64, or exfiltration; throttled to once/24h with an opt-out env var. Trigger: judged "end session", "整理文件", "status where are we" as loads and "what's the current git status", "set up a new React project with Vite" as non-loads (5/5, though "status"/"setup" carry mild keyword-collision risk). Output test: wrote /private/tmp/.../scratchpad/baseline.md (9-line prose, 0 headers, no per-task fields) vs skill-log.md following the Session Log Format strictly (29 lines, 7 canonical headers, each pending task with the mandated Owner|Deadline|Dependencies scaffold, plus a surfaced CLAUDE.md candidate the baseline omitted).
Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works
- Installs cleanly 5/5
- Triggers reliably 5/5
- Output vs. baseline 7/10
- Docs & honesty 5/5
What Project Butler does
Maintains a standardized project-memory stack (CLAUDE.md, PROJECT.md wiki, session-handoff, TODO, session logs, update log, docs index, profile config) so coding agents resume projects without re-explanation. Triggers on /project-butler, "end session"/"收工", "continue"/"接着上次", "status", "organize files"/"整理文件", "sync wiki", "review claude", and language switching. Includes a daily git-based version freshness check that offers to self-update.
How to install Project Butler
git clone --depth 1 https://github.com/JamesShi96/project-butler.git ~/.claude/skills/project-butler
# SKILL.md sits at the repo root, so clone straight into the skill dir.
# No API keys or external deps. Optional: the skill runs scripts/check-update.sh
# (a daily `git fetch` against its own repo) to offer self-updates; set
# PROJECT_BUTLER_NO_UPDATE_CHECK=1 to silence it. Works with Claude Code, and
# Cursor (.cursor/rules) / Codex (AGENTS.md) via templates it can generate.
Skills live in ~/.claude/skills/ (global) or .claude/skills/
(per-project). Restart Claude Code after installing.
Commands — how to trigger Project Butler
-
/project-butlerCross-session project memory for AI coding agents: session logs, handoff, TODO, wiki, status
It also activates on plain-language prompts like these:
-
Set up project memory for this new project -
Close out this session and update the log -
Organize the files in this project folder
Frequently asked questions
- Is the Project Butler skill free?
- Yes. The skill itself is free from JamesShi96/project-butler. SkillProof publishes the install command and an independent test verdict at no cost.
- Does Project Butler work with Claude Code?
- We tested it with Claude Code 2.x (agent harness) on Jul 31, 2026. Verdict: Tested · Works. Fetched root SKILL.md; spot-checked references/update-check.md, scripts/check-update.sh, references/file-templates.md, references/project-profile-system.md (all HTTP 200). Read check-update.sh fully: it only does git fetch/pull on its own dir and greps CLAUDE.md for a language token — no curl|sh, base64, or exfiltration; throttled to once/24h with an opt-out env var. Trigger: judged "end session", "整理文件", "status where are we" as loads and "what's the current git status", "set up a new React project with Vite" as non-loads (5/5, though "status"/"setup" carry mild keyword-collision risk). Output test: wrote /private/tmp/.../scratchpad/baseline.md (9-line prose, 0 headers, no per-task fields) vs skill-log.md following the Session Log Format strictly (29 lines, 7 canonical headers, each pending task with the mandated Owner|Deadline|Dependencies scaffold, plus a surfaced CLAUDE.md candidate the baseline omitted).
- What is the Project Butler SkillProof Score?
- 8.8/10 — installs cleanly 5/5, triggers reliably 5/5, output vs. baseline 7/10, docs & honesty 5/5.
- How do I install Project Butler?
- Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
- Can I use Project Butler with Cursor, Copilot, Gemini CLI, Codex or other AI tools?
- The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.