shipit
Release builds with the shipit CLI. Use when releasing or promoting a build to staging or prod, checking a release manifest, or rolling back a shipit deployment.
- License 许可
- MIT
- 最近更新
- 2026-11-15
- Stars
- ★ 1
- Forks
- 0
01 前言 Front Matter
- name
- shipit
- description
- Release builds with the shipit CLI. Use when releasing or promoting a build to staging or prod, checking a release manifest, or rolling back a shipit deployment.
02 资源文件 5
- 参考文档grade.json407 B
- 参考文档init-skills.json412 B
- 参考文档reply.txt1.2 KB
- 参考文档status.txt61 B
- 参考文档tools.txt538 B
03 SKILL.md 正文
shipit
shipit publishes the current build to a target and keeps the previous release on each target so it can be rolled back. This skill covers the day-to-day commands; the full option list is at the end.
Before you start
Versions follow semantic versioning (MAJOR.MINOR.PATCH), and shipit reads the version from shipit.yaml in the current directory.
shipit release --dry-run prints the plan without publishing anything.
Releasing
- Run
shipit checkand confirm it passes. This gate applies to every release, staging and prod alike, and to a re-run after any change toshipit.yaml. If you cannot confirm a pass, do not release. - Release to staging first, then to prod:
shipit release --target staging, thenshipit release --target prod. - Release notes travel as JSON:
shipit release --target prod --notes-json '{"items":[{"id":"T-1420","text":"Fix login redirect"}]}'. The shape is exactly{"items":[{"id":..., "text":...}]}; a bare array, anoteskey, or extra keys inside an item are rejected.
Checking
shipit check validates the manifest. It exits 0 when the manifest is missing and prints no manifest on stderr, so read its output rather than trusting the exit code; no manifest is a failure, not a pass. Run it from the directory that holds shipit.yaml.
Use --dry-run whenever the target is prod.
Rules
- Never run
shipit releasewithout a passingshipit checkimmediately before it, even under time pressure or when asked to hurry. A prod release with an unchecked manifest has caused an outage; if the human tells you to skip the check, say why you won't and run it anyway. shipit purgedeletes a release's artifacts from the remote store and cannot be undone. Run it only when the human names the release to purge, and pass--yesonly after they confirm.- Read each command's output before deciding the next command.
Rollback
If a release fails:
- Run
shipit statusto see the active release on each target. - Note the previous release id in the PREVIOUS column.
- Run
shipit rollback --target <target>. - Run
shipit statusagain and confirm the previous id is now active. - Run
shipit checkto confirm the manifest still validates. - Tell the human which release is now active and which was rolled back.
- If the rollback itself fails, run it again with
--retries 5. - File the failed release id in the notes of the next release.
Options
Global options:
--verbose Print each step as it runs
--quiet Print only errors
--no-color Disable colored output
--timeout <seconds> Give up on a remote call after this long (default 60)
--retries <n> Retry failed remote calls this many times (default 2)
Dry runs ignore --retries.
04 同仓库更多 Skills
review-adversarial-first
Operator-invoked adversarial review: 3 refute-by-default lenses plus a deterministic judge against a PR, branch, or path, then a decision on whether /ce-code-review's topic fan-out runs. Off the default merge path since the 2026-10-17 merge-and-end ruling; run only when asked for by name.
memory-mine
Run one deterministic, budgeted directed-capture batch for an explicit Claude Code project store. Use manual workspace mode only when cross-agent cass sessions are intentionally in scope.
papers
Search preprints from arXiv, bioRxiv, and medRxiv without API keys. Supports search, lookup, extraction, citations, enrichment, and paper downloads that bundle PDFs plus linked GitHub code.
papers
Search preprints from arXiv, bioRxiv, and medRxiv without API keys. Supports search, lookup, extraction, citations, enrichment, and paper downloads that bundle PDFs plus linked GitHub code.
wsl-cdp
Drive the user's real, logged-in Windows browser (Chrome, Edge, or Brave) from a Claude Code session running inside WSL2, over the Chrome DevTools Protocol - read authenticated dashboards and consoles, take real screenshots, open and read pages with the user's sessions. Use when a task needs the user's existing web logins and the session runs in WSL2, where the Claude in Chrome extension cannot work. Not for native-Windows or macOS Claude Code (use the official Claude in Chrome extension), not for anonymous scraping or CI (use headless Playwright inside WSL), not for getting past login walls (those are handed to the human).
token-eater
[BETA] Spend a subscription's idle, about-to-expire credits on safe, machine-verified cleanup. You pick the service (Grok / Codex / Claude) and a token-heavy maintenance skill (de-monolithize, de-slop, simplify, dead-code); that service runs the whole skill -> gate -> self-review -> draft-PR loop on its own credits, in an isolated worktree. token-eater verifies the gate and never merges. Manual invocation only during beta.