Issue Lifecycle¶
Most work in bipartite follows one path: draft an issue, check it, file it, implement it on a branch, gate the PR, land it, and turn its follow-ups into the next issues.
Each step is a skill.
This page gives only the order.
Each skill's own SKILL.md (in skills/) says what it does and how, and the description: line at its top is the one-line summary.
flowchart LR
D[Draft] --> C[Check & file] --> W[Work] --> G[Gate] --> L[Land]
L -. DEFERRED follow-ups .-> D
1. Draft¶
/bip-issue-nextdraftsISSUE-<slug>.mdin the repo root from a PR's follow-ups, a focus file, or the current conversation, then hands it to/bip-issue-check./bip-issue-iterateworks through a draft paragraph by paragraph with you over many rounds, for one that needs to get smaller and more correct first./bip-remove-metaspeakcuts reasoning-out-loud residue from a draft that has been through several rounds.
ISSUE-*.md files are drafts and are never committed.
2. Check and file¶
/bip-issue-check ISSUE-<slug>.mdreviews the draft for implementation-readiness, fixes gaps, and submits it via/bip-issue-file./bip-issue-update Nre-checks an issue that is already filed, which may have drifted from the code since.
3. Work¶
Pick one by where you want the work to run:
| Where | How |
|---|---|
| This session | /bip-issue-work N |
| A new tmux window with the issue loaded | bip spawn org/repo#N (Workflow Coordination) |
| An EPIC's clone pool, autonomously | /bip-conductor-spawn, run from a /bip-conductor session (see the README's orchestration section) |
The worker branches (per-issue worktrees are optional: see Worktree Layout), implements, and opens a PR with Closes #N.
Work it can't fold in goes in the PR body's DEFERRED section.
4. Gate¶
/bip-pr-checkis the quick readiness check (clean tree, description, body)./bip-pr-reviewis the full pre-merge review. It uses the repo'sPRE-MERGE-CHECKLIST.mdif there is one./bip-comment-checkchecks a reviewer's comment against the code before you act on it.
5. Land¶
/bip-pr-land squash-merges the PR, deletes its branches, and returns to an up-to-date main.
Follow-ups from DEFERRED go back to step 1 via /bip-issue-next <PR>.
Across context resets¶
A session that fills its context runs /bip-tuckin before /clear and /bip-continue after (Continuation prompt).
A fresh session dropped into a spawned clone that already has a PR or a stalled worker starts with /bip-spawn-resume.