agentsclimarketplace

Figloops go plan ack

Skill tyburrowbridge/figloops/skills/figloops-go-plan-ack

Plan-ack phase — render plan frame in Figma, anchor comment threads, refresh on subsequent runsFrom its SKILL.md

Install
npx -y skills add tyburrowbridge/figloops --skill figloops-go-plan-ack

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

3.5 KB, as published. Nobody here has run it

Setup

Resolve FIGLOOPS_PLUGIN_DIR from env or .env. If unset, abort. Resolve FIGMA_TOKEN from env or .env. Scripts: cd "<CONSUMING_REPO>" && "<PLUGIN_DIR>/node_modules/.bin/tsx" "<PLUGIN_DIR>/scripts/<name>.ts" <args>. Always double-quote paths.

Style

Status table per round. AskUserQuestion options carry no descriptions.

Errors

TS exits non-zero → relay stderr verbatim, don't advance. MCP failure → prompt user to retry; don't advance.


Handler

  1. Mark [FIGLOOPS] Ack plan in Figma as in_progress.

  2. Read feedback/state.json and figloops.config.json to get figma.fileKey and current round.

  3. Mode detection. If state.rounds[currentRound].planFrame is unset → render+anchor mode (step 4). Else → refresh mode (step 7).


Render + anchor mode

  1. Render frame. Run:

    PAYLOAD=$("<PLUGIN_DIR>/node_modules/.bin/tsx" "<PLUGIN_DIR>/scripts/render-plan-frame.ts" "feedback/state.json" <round>)
    

    Then invoke use_figma with fileKey: <fileKey>, code: "$PAYLOAD", description: "Render Round <N> plan frame". Capture the returned JSON: {pageId, frameId, frameName, rows: [{itemId, index, nodeId}]}.

  2. Persist frame result + assign rowIndex. Read state.json, set rounds[round].planFrame = {pageId, frameId, frameName}, set rowIndex on each plan item from rows. Write back.

  3. Anchor threads. Run:

    "<PLUGIN_DIR>/node_modules/.bin/tsx" "<PLUGIN_DIR>/scripts/anchor-plan-threads.ts" "feedback/state.json" <round> "<fileKey>" "$FIGMA_TOKEN"
    

    Regenerate snapshot.

    Print:

    📋 Round <N> plan rendered in Figma — <count> items anchored.
    
    Open the Figma file and act on each item:
    - ✓ resolve the thread when you've shipped it
    - reply `/skip` to mark won't do
    - leave it open to discuss
    
    When you're done (or want to check in), re-run `/figloops:go`.
    

    Exit. Do not advance.


Refresh mode

  1. Pull thread states. Run:

    DELTA=$("<PLUGIN_DIR>/node_modules/.bin/tsx" "<PLUGIN_DIR>/scripts/pull-plan-states.ts" "feedback/state.json" <round> "<fileKey>" "$FIGMA_TOKEN")
    

    Parse the JSON: {statuses: {itemId → status}, delta: {shipped: [], wontdo: [], pending: [], removed: []}}. The script has already written the new statuses back to state.json.

  2. Recovery check — plan frame deleted. If any item now has status: 'removed' AND it was previously not 'removed', run get_metadata on planFrame.frameId to confirm. If frame is missing too, use AskUserQuestion:

    question: "Plan frame for Round <N> is missing. Re-render?"
    header: "Recovery"
    options:
      - label: "Re-render frame and threads  (Recommended)"
      - label: "Cancel — leave state as-is"
    
    • Re-render: clear planFrame and clear commentId/botReplyId on every item, then re-enter render+anchor mode (step 4).
    • Cancel: print the delta and exit.
  3. Print delta.

    🔄 Refreshed Round <N> plan
    
    + <N> newly shipped (<#list>)
    + <N> newly skipped (<#list>)
    = <N> still pending
    

    Regenerate snapshot.

  4. Advance check. Count items by status from the new state.

    • If any item is pending → exit. Tell user to keep working in Figma.
    • Else (all shipped / wontdo / removed) → mark task completed, run advance-phase.ts close, invoke skill figloops-go-close.

Keep looking

Skills are one crate of 608,666. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.