bridge-harness
CLAUDE.md4 starsChanged 6 months agoMITView on GitHub
- Before every commit, bump the version field in both package.json files: - package.json — main package @cocodrino/bridge-harness - packages/bridge-harness-pi/package.json — @cocodrino/bridge-harness-pi - Both versions MUST always match. The two packages are released in lockstep — never commit with stale or mismatched versions. - Bump the patch level by default during 0.x. - The Claude Code side runs from the local repo dist/ (see ~/.claude.json), so run npm run build after any change under src/ (MCP server / shared) before…
In this repository5 other files it gives its agents.
- openspec-apply-changeSkill
- openspec-archive-changeSkill
- openspec-exploreSkill
- openspec-proposeSkill
- agent-bridgeSkill
The file
Files
bridge-harness — Project Instructions
Versioning (MANDATORY)
- Before every commit, bump the
versionfield in bothpackage.jsonfiles:package.json— main package@cocodrino/bridge-harnesspackages/bridge-harness-pi/package.json—@cocodrino/bridge-harness-pi
- Both versions MUST always match. The two packages are released in lockstep — never commit with stale or mismatched versions.
- Bump the patch level by default during
0.x.
Build & publish
- The Claude Code side runs from the local repo
dist/(see~/.claude.json), so runnpm run buildafter any change undersrc/(MCP server / shared) before committing. - The Pi extension ships
src/directly (no build needed to publish), but runtsc --noEmitinpackages/bridge-harness-pi/to typecheck. npm publishrequires a 2FA OTP.
Architecture notes
project= the NATS namespace, derived from the git worktree root (orBRIDGE_PROJECT). Everything keys off it: subjects, registry, presence, and the default lobby room.- Both agents auto-join the room named after the project.
use_bridgeswitches the namespace at runtime. - Running MCP/extension processes hold code in memory — rebuilding
dist/or republishing only affects newly started processes. Restart sessions to pick up changes.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

