blitzos

agent
Security Audit
Fail
Health Pass
  • License — License: MIT
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Community trust — 12 GitHub stars
Code Fail
  • rm -rf — Recursive force deletion command in skill/scripts/build-monorepo.sh
  • rm -rf — Recursive force deletion command in skill/scripts/scan.sh
  • rm -rf — Recursive force deletion command in tests/build-thin-repo.sh
Permissions Pass
  • Permissions — No dangerous permissions requested

No AI report is available for this listing yet.

SUMMARY

BlitzOS — warm-start context for your Claude Code / Codex cloud sessions. One command builds a company-context repo so cloud sessions boot knowing your whole codebase. If you want to set this up for your company, follow the link below.

README.md

BlitzOS

Claude/Codex Cloud VMs that boot knowing your whole codebase

Website
Discord
License


Your Claude Code cloud sessions, booted knowing your whole codebase. blitzos is a free, open-source skill that builds a thin private context repo so cloud sessions start warm: your repos, conventions, and a running work log, without copying code, history, or secrets.

Early and rough in places.

Managed BlitzOS

want this for your team, with a credential vault, per-task scoped access, and your own VPC? Join the waitlist

Quickstart

Needs Claude Code, Git, gh (authenticated), jq, and Node 18+.

git clone https://github.com/blitzdotdev/blitzos.git
cd blitzos && ./install.sh
claude "set up blitzos"

The skill scans your repos, opens a localhost wizard, shows a context draft for your approval, and creates a private context repo. Then in claude.ai/code, select that repo alongside your work repos and start a session.

What you get

  • Warm multi-repo context — sessions start with your repo map, conventions, and recent work log.
  • Your connectors — Claude can use Linear, Slack, or Gmail when they're connected to your account.
  • A real cloud VM — install, build, and test across your repos.
  • A session log — short records in sessions/ carry work forward between sessions.

The default Trusted network is enough — git and connectors use separate proxies, and registries are reachable. Only widen it (prefer Custom over Full) if a task must reach an outside host.

Not yet

Claude-only (Codex coming). The free tier has no managed credentials, deploy, or per-task scoping — that's Managed BlitzOS. Pushing the session log back from cloud is experimental.

Troubleshooting

Claude can "add" files from my repo but 404s when I ask it to read the repo — even with full access.
You're likely using the GitHub connector in regular Claude chat, which only attaches individual files. To have Claude actually read and work across a repo, use Claude Code on the webclaude.ai/code: connect GitHub there, then select the repo when you start a session. The session clones the whole repo into its environment and reads files natively — no 404. The chat connector and the coding session are different surfaces; this skill targets the coding session.

Still 404ing in claude.ai/code? Make sure the Claude GitHub App is authorized on that specific repo (GitHub → Settings → Applications → Claude). Org-owned and private repos often need to be granted individually, even if your account already has "full access."

License

MIT — see LICENSE.

Reviews (0)

No results found