From jahns-workflow
This skill should be used when the user runs "/jahns-workflow:round", says to "close the round", "wrap up this round", "finish the work cycle", or when an autonomous work round (implement → verify → push) reaches its end and the project CLAUDE.md mandates round closeout. Updates the task registry, PROGRESS, roadmap, SSOT views, and produces the external review packet.
How this skill is triggered — by the user, by Claude, or both
Slash command
/jahns-workflow:round [round-slug] e.g. lstream-seams[round-slug] e.g. lstream-seamsThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Close the current work round: bring the task registry up to date, record the round in
Close the current work round: bring the task registry up to date, record the round in PROGRESS, refresh generated views, and emit a paste-ready external review packet.
Requires an initialized project (.jahns-workflow.yml). If missing, stop and point the user
at /jahns-workflow:init. Plugin root = two directories above this skill's base directory.
<today YYYY-MM-DD>-<slug>: slug from the argument if given, else derive a short one from the
round's dominant theme. Check PROGRESS for an existing entry with the same id (extend it
rather than duplicating).
Update tasks.yaml to match reality (the guard hook auto-validates and regenerates
ROADMAP.md on save — if it reports violations, fix them immediately):
done (a gate/... task is done only if the bar actually passed — link evidence in PROGRESS).round: <round-id> on every task worked this round.<type>/<slug> IDs + explanatory titles; set anchor: to the governing SSOT §-anchor when known — audits scope by it). Unresolved questions for the user become decision/... tasks.blocked states from deps.Skip if the config has no ssot:. Otherwise:
uv run <plugin-root>/scripts/jw_ssot.py check .
# stale (exit 3) → regenerate:
uv run <plugin-root>/scripts/jw_ssot.py split . && uv run <plugin-root>/scripts/jw_ssot.py digest .
Then measure the round's SSOT churn: git diff <watermark> HEAD --numstat -- <ssot>, where
the watermark is state.last_round_commit from .jahns-workflow.yml; if null, fall back to
the most recent commit whose subject starts with docs(round):; if neither exists, skip the
measurement and note that. If added+deleted > 100 lines, apply the bulk-edit quarantine
rule: state prominently in the report and in PROGRESS that /jahns-workflow:audit must run
on the changed sections before dependent work consumes them. Finally, update
state.last_round_commit to the current HEAD.
Append an entry from <plugin-root>/templates/progress-entry.md (content in the user's
configured language). Then archive: move dated sections from months before the current one
into docs/progress/<YYYY-MM>.md (mechanical cut-paste, newest-first preserved), leaving
PROGRESS.md with the current month + the header pointers.
Generate <reviews_dir>/<round-id>-request.md from <plugin-root>/templates/review-request.md.
External reviewers typically browse the repo directly (e.g. ChatGPT's GitHub connector), so:
record the pushed HEAD hash in the packet (verify the round's commits are actually pushed —
git status -sb; if not, note that the user must push before requesting review), and prefer
pointers (file paths, §-anchors, commit hashes) over inlined diffs — inline only small
load-bearing snippets, or full diffs/pseudocode if the reviewer has no repo access. State
every load-bearing claim falsifiably and list the test ladder's known blind spots.
Report in the user's configured language: shipped tasks (id — title), registry/roadmap state,
SSOT churn (+ quarantine flag if triggered), where the review packet is, and a suggested
commit message (docs(round): close <round-id>). Remind: paste the packet to the external
reviewer and ingest the reply with /jahns-workflow:review <round-id>. Do not commit unless
the project's conventions say rounds end in a commit and the user has authorized committing.
Guides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.
npx claudepluginhub dev-jahn/jahns-cc-marketplace --plugin jahns-workflow