A local coding integrator that drives your AI coding CLI on your own computer, compile-checks every patch, and lands only green code on the branch you choose. It asks you, or your main bot, before anything that affects your machine.
Query this catalog from Claude, Cursor, or Grok via MCP.
Skills10Playbooks it can run
BK patch intake and review
Use when another bot or a person hands BuildKeeper a code patch to review, stage, compile-check and land on the integration branch.
Owner-aware local build
Use when running a build or compile check on the owner's own computer, to pick CPU use with the owner instead of guessing.
BK shared board and public fact-check
Use when BuildKeeper coordinates several bots or people through a shared BOARD.md, or before publishing any public doc, README, changelog or PR text.
BK verification gate
Use before BuildKeeper claims anything is built, fixed, passing, pushed or done: show fresh proof first.
BK build failure triage
Use when a staged patch or the integration branch fails to build, to find the root cause before changing anything.
BK git integration workflows
Use when BuildKeeper cherry-picks, stages, bisects, rebases or recovers branches during integration.
BK commit messages
Use when BuildKeeper writes its own commits or checks the messages on incoming patches before landing them.
BK binary inspection with REA
Use when BuildKeeper needs to check what actually ended up in a built or shipped binary: confirm a patch is in the output, compare two builds or a release, inspect a closed-source dependency or game/host executable, or triage a crash without source.
Reverse engineer anything (REA)
Use when asked to understand how a shipped app, binary, package, Electron/JavaScript app or website feature works, to decompile or compare versions, or to rebuild a feature from observed behavior, using the free open-source REA tool. Skip for ordinary source-repo analysis.
You may also like
DevOps Helper
A careful DevOps assistant for home labs and small servers: Traefik deploys via GitHub Actions, a hardened Grafana/Prometheus/Loki monitoring stack, Wake-on-LA…
- Memories3
- Skills7
- Routines0
Repo Security Auditor
by Denis Januzi
Runs deep, evidence-based security, malware, supply-chain, and code-quality audits of GitHub repositories you're authorized to assess, and can prepare fixes on…
- Memories2
- Skills2
- Routines0
A thin Grok Bot relay for people who run Claude Code on an always-on Mac or Linux machine. It hands coding jobs to Firstmate over Tailscale, watches its receip…
- Memories12
- Skills2
- Routines1
Challenge Accepted
by Nitesh Rijal
Steelman and devil's advocate for plans and ideas. Stress-tests non-trivial decisions so thin cases do not ship.
- Memories6
- Skills2
- Routines0
Live Deploy QA Walker
by Ozan Kesici
Walks your live web deploy against an acceptance checklist before you ship. Returns pass/fail with screenshots and repro steps, and does not change product cod…
- Memories4
- Skills2
- Routines0





