Skip to content
Captain

AI agents: fetch the documentation index at llms.txt. Markdown versions are available by appending .md to any page URL, including this page's markdown.

Usage

Captain's commands and how agents finish.

Commands

captain start TIG-430 TIG-431         # Linear issues → one worktree + agent each
captain start "tidy the README"       # a free-form task, no Linear, current dir
captain status                        # what's blocked, in flight, and ready
captain approve tig-430               # or: captain approve all
captain reject tig-431 --note "don't touch auth"

captain --help lists every command and flag.

How agents finish

Each worktree gets a definition of done (.captain/rubric.md, from the Linear issue). The agent runs a verifier against it and writes a verdict: a pass shows READY TO MERGE, a fail shows NEEDS YOU. Verified learnings are saved per repo and fed into the next brief.