agentleFS
fifersheepfifersheep/flow/skills/flow-wrap/SKILL.md

flow-wrap

Skill0 starsChanged 3 months agoApache-2.0View on GitHub

Flow phase 8 (wind-down; also the lite-mode wrap). Normally invoked via /flow. Learnings capture + marks the task complete ON THE BRANCH, then merge + cleanup — so the completion marker rides in the original PR, never a follow-up.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Ask approve-merge / request-changes.

    line 14

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Flow — wrap

Lite mode entry (arriving from code-review)

Gate 2 hasn't run yet — run it first, in miniature:

  1. Append a ## Quick report section to workflow.md: 3–6 lines — what changed, verify output (pasted), deviations if any.
  2. Open the PR (or present the local merge) as in flow-report step 2.
  3. Ask approve-merge / request-changes. On approval append - YYYY-MM-DD gate 2: merge approved by user, THEN set phase: wind-down.

Wind-down (both modes)

Order matters: everything that touches a repo file — learnings AND the phase: complete marker — is committed to the task branch BEFORE the merge, so it all rides in the original PR. Never open a separate "mark complete" follow-up PR.

  1. Learnings capture. Review the task for durable lessons: a gotcha hit twice, a correction the user made, a convention discovered. For each, propose a concrete diff to CLAUDE.md, a domain skill, or memory — ONE at a time, user approves each. No approval → no edit. Zero lessons is a valid outcome; never invent one. Repo-file learnings (e.g. CLAUDE.md) are committed to the branch; memory lives outside the repo.
  2. Close the record on the branch. Set phase: complete, final updated bump, and commit ALL wind-down repo changes (approved learnings + this marker) to the task branch. This is the last commit before the merge, so the PR carries the finished task record. (If the user deferred the merge, still commit this, but leave phase: wind-down until they actually merge.)
  3. Merge per the gate-2 decision (merge PR / local merge to the default branch) — the PR now includes the learnings and the completion marker. If the user chose not to merge yet, skip the merge and go to cleanup. Once the merge has landed, move the ticket to the merged status per the plugin's templates/tracker-sync.md and record tracker-synced. No merge → no move; the ticket stays where it is.
  4. Cleanup. git status must be clean; remove the worktree if one was created; delete the merged branch if the user agrees; verify nothing uncommitted is stranded. The task folder stays forever as the record.
  5. Tracker drift check. With a tracker configured and a merge landed, tracker-synced must now equal the merged status. Anything else — an earlier sync that failed, a status renamed mid-task — is reported to the user as a one-line "ticket <id> is at <X>, expected <Y> — move it by hand" at wind-down, alongside the artifact paths. This is the last chance to catch it, so never skip it and never silently "fix" it by re-writing tracker-synced without a successful move.

Announce completion with the artifact paths and any drift note.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.