plan-update

Étoiles10

Forks0

Mis à jour22 juillet 2026 à 18:31

Edit a plan's frontmatter (status, assigned agent, and other fields) via plan_keeper_cli.py file-meta set. Use when the user asks to change the agent on a plan, promote a plan to todo, mark a plan as ready, set a plan's status, or edit plan frontmatter.

Installation

Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.

Source

paulbaranowski

paulbaranowski/wild-horses

Ouvrir le dépôt GitHub Voir les dépôts du créateur

Téléchargement

Métiers associésSOC

Basé sur la classification professionnelle SOC

Développeurs de logicielsProfessions informatiques et mathématiques·SOC 15-1252

Explorateur de fichiers

2 fichiers

SKILL.md

readonly

Plus depuis ce dépôt

même dépôt

summary-writer

paulbaranowski/wild-horses

Write pull-request descriptions and titles that lead with the one structural idea, not a file-by-file changelog. Use when about to write or revise a PR description or title (before `gh pr create`/`gh pr edit`, or when a repo squashes with the PR body as the commit message), or when an existing description reads like a changelog and needs rewriting.

2026-07-3110

autonomous

paulbaranowski/wild-horses

Autonomously take an issue/ticket or a plan file from a link (or path) to an opened pull request, with no human in the loop. Hand it a Linear/GitHub/other issue URL, a path to a plan/spec file (e.g. ~/plans/repo/foo.md), or a plan already read into the conversation, and it decides everything itself: implements, tests, simplifies the diff, runs a bounded reasoning-gaps review (critical findings only, via harness agent prompts), reviews via wild-pr:review (falling back to an independent sub-agent) to convergence, and opens a PR following the target repo's own conventions. Use when the user says work this issue autonomously, take this ticket end-to-end, do this AFK, pastes an issue link, or points it at a plan file and asks you to just build it. Ships an autonomy contract (never stop to ask) plus a 10-rule code-style bar.

2026-07-2210

cleanup-worktrees

paulbaranowski/wild-horses

Reclaim disk space from stale git worktrees: scans configured roots, classifies each as cleanable (merged PR, upstream gone, merged into the default branch, stale) or skipped (dirty, locked, unpushed), shows the cleanable set grouped by reason with sizes, and removes the ones you pick. Use when the user asks to clean up git worktrees, delete merged worktrees, prune stale worktrees, free disk space from worktrees, or list cleanable worktrees.

2026-07-2210

codepath-mapper

paulbaranowski/wild-horses

Scan a codebase and produce architecture.json + codepaths.json describing components, edges, and traced codepaths. Renders architecture.html at the end. Use when the user says "map my app", "build the architecture viz", "scan my codepaths", "map the X flow", etc. Pass `--render-only` to skip all mapping work and just re-bake architecture.html from existing JSONs — use when the user says "re-render the visualizer", "rebuild the viewer HTML", "regenerate architecture.html", or after a template/CSS edit.

2026-07-2210

codepath-visualizer

paulbaranowski/wild-horses

Open docs/codepaths/architecture.html in the browser. With --select, blocks on a user pick and returns the chosen codepath JSON to the calling agent. Run after /codepath-mapper has produced the HTML.

2026-07-2210

task-list-builder

paulbaranowski/wild-horses

Build or rewrite a structured task list (JSON + paired markdown report) matching the harness task-list schema. Accepts free-form text, an existing reasoning-gaps/feedback-blockers report, an existing JSON task file (in-place rewrite), or recent conversation context. Use when the user says "build the task list using task-list-builder", "rewrite the plan file using task-list-builder", or otherwise asks to convert or update a chunk of work into the harness task-list format.

2026-07-2210

name	plan-update
description	Edit a plan's frontmatter (status, assigned agent, and other fields) via plan_keeper_cli.py file-meta set. Use when the user asks to change the agent on a plan, promote a plan to todo, mark a plan as ready, set a plan's status, or edit plan frontmatter.

plan-update

Edit the frontmatter of an existing plan in ~/plans/<repo>/. The bundled plan_keeper_cli.py file-meta set does the atomic write — one self-documenting flag per field; this skill's job is to identify the plan, identify which field(s) to change, and route to the CLI after user confirmation.

Quick reference

Target: ~/plans/<repo>/<filename> (active state — not done/ or deferred/).
Field → flag: Agent → --agent, Status → --status, Kind → --kind, Completed on → --completed-on, Plan-keeper Ticket → --plankeeper-ticket, Linear Ticket → --linear-ticket, Jira Ticket → --jira-ticket. (--ticket is not a value flag — it locates a plan by any of its id fields, like push; write an id with the matching per-system flag.)
--status is lifecycle-aware: active states (backlog/todo/in-progress/in-review) rewrite in place, but --status done/--status deferred relocate the plan into done//deferred/ (and done stamps Completed on) — exactly what plan-done does. Prefer plan-done for completing a plan; reach here for done/deferred only when editing other fields in the same breath.
Status vocabulary: backlog (default; fetched but not dispatched — confirm via crew status <id>), todo (the status gate for dispatch — but not sufficient alone: groundcrew also requires an Agent: tag and a registered repo, see ../../groundcrew/README.md), in-progress (set by groundcrew's markInProgress hook), in-review (set by groundcrew's markInReview hook when the PR opens), done (set by plan-done when archiving). The middle values (in-progress, in-review, done) are normally written by the system — set them by hand only if you know why.
Kind vocabulary: idea / prd / reqs / design / spec / exec-plan — the document type, validated against this closed set (see ../../plan-kinds.md, including its "Choosing the right Kind" decision procedure). Set by plan-save; correct it here if it was inferred wrong. Changing --kind renames the file's --<kind> segment to match (frontmatter stays the source of truth), so the path changes and the CLI prints the new one.
Common edits:
- Promote: --status todo sets the status gate; groundcrew won't dispatch until the plan also has an Agent: tag, so pair it with --agent claude (or queue via plan-crew, which stamps the Agent for you).
- Change model: --agent codex.
- Reset: --status backlog.
- Reclassify: --kind exec-plan (changes how plan-do routes the plan, and renames the --<kind> filename segment to match).
Confirmation: required before any mutation.

Procedure

1. Identify the plan

If the user just referenced a specific file ("this plan", "the one I just saved"), use that.

Otherwise, list active plans:

python3 "${CLAUDE_PLUGIN_ROOT}/scripts/plan_keeper_cli.py" list

Run this command fresh every time you reach this step — including on a re-invocation later in the same conversation. Never reprint an earlier listing from memory: plans get saved, archived, or change status between turns, so a cached list can be stale. The numbered list you show must come from the output you just ran.

Present numbered to the user; they pick.

2. Identify the field(s) to change

Match the user's invocation:

"promote to todo" / "mark ready" / "set status to todo" → --status todo
"change agent to codex" / "use codex" → --agent codex
"back to backlog" / "reset" → --status backlog
"it's actually a spec/design/reqs/prd/idea/exec-plan" / "reclassify" → --kind <value>
"set the linear ticket to ENG-123" → --linear-ticket ENG-123 (use --jira-ticket / --plankeeper-ticket for the other systems)
Anything else: ask which field, which value.

Multiple fields → pass multiple flags in one call.

3. Confirm

About to update <filename>:

<Key>: <old> → <new>

(repeat per field)

Proceed?

Do not skip — even for "obvious" promotions.

4. Run the update

python3 "${CLAUDE_PLUGIN_ROOT}/scripts/plan_keeper_cli.py" file-meta set \
  --file "$HOME/plans/<repo>/<filename>" \
  --<flag> <value>

Pass additional field flags in the same call as needed (e.g. --status todo --agent codex). --ticket <id> is an alternative to --file for locating the plan — it finds the plan by any of its id fields (Plan-keeper Ticket / Linear Ticket / Jira Ticket) across all repos (exactly one of --file/--ticket is required); don't confuse it with the per-system value flags (--plankeeper-ticket / --linear-ticket / --jira-ticket), which write an id.

5. Confirm the result

The command in step 4 prints the plan's path on stdout. That path can differ from the input when the edit relocates the file: --status done/--status deferred moves it into done//deferred/, and --kind re-stamps the --<kind> filename segment. Use the printed path here, not the filename you passed in:

python3 "${CLAUDE_PLUGIN_ROOT}/scripts/plan_keeper_cli.py" file-meta get \
  --file "<path the set command printed>"

Edge cases

Plan has no frontmatter (or a partial block) — the CLI adopts it: it stamps plan-save's defaults (Status/Created/minted Plan-keeper Ticket), backfilling only the absent managed fields, before applying your edit, and prints adopted unmanaged plan ... on stderr. No re-save step is needed. Only genuinely malformed frontmatter (a --- opener with no closing ---) is still rejected — there the CLI exits 5 with "malformed frontmatter" and leaves the file untouched.
Field has no flag — set only edits the known fields (the field→flag list above). A field with no flag isn't editable here; for a genuinely new custom field, point the user at the spec for extending _FRONTMATTER_FIELDS.
Invalid --kind / --completed-on — CLI exits 2 (Kind must be in the closed set; date must be YYYY-MM-DD) and the file is left untouched (inputs are validated before any write).

Notes

Setting an active status (backlog/todo/in-progress/in-review) only rewrites frontmatter in place. Setting --status done/--status deferred relocates the file into done//deferred/ (the CLI's lifecycle behavior) — same as plan-done. Always confirm before invoking, since that move is destructive; prefer plan-done for completing a plan.
This skill never creates files — that's plan-save's job.
The mutation is atomic (tmp file + fsync + os.replace) so a crash mid-update can't corrupt the plan.
For promoting many plans at once, or browsing the queue across all repos, use the plan-crew skill — it's the cross-repo, multi-select counterpart. plan-update stays the targeted single-plan / current-repo editor (and the way to set Agent and the per-system ticket fields, etc.).