enrich

Enrich Thin Frames

1. Parse arguments

Extract fileKey from $ARGUMENTS:

• If it looks like a Figma URL, extract the fileKey segment: https://www.figma.com/design/<fileKey>/...
• If it's a bare fileKey (no slashes), use it directly

2. Check prerequisites

bash $CLAUDE_PLUGIN_ROOT/scripts/auth.sh status

If it fails, tell the user to run bash $CLAUDE_PLUGIN_ROOT/scripts/auth.sh set and stop.

Verify the fileKey has local data:

ls ~/.figma-differ/<fileKey>/

If missing, tell the user to run /figma-differ:track <url> first.

3. Orchestration

TaskCreate("Scan thin frames",   activeForm: "Detecting frames with shallow data...")
TaskCreate("Re-fetch & enrich",  activeForm: "Fetching full node JSON for thin frames...")
TaskCreate("Rebuild frame docs", activeForm: "Regenerating frame.md documents...")
TaskCreate("Update search index", activeForm: "Re-indexing enriched frames in QMD...")

Task lifecycle:

1. Create ALL tasks upfront (pending)
2. Before each phase: TaskUpdate(taskId, status: "in_progress")
3. Dispatch Agent(model: "haiku") — agent reports only counts, never raw output
4. After agent returns: TaskUpdate(taskId, status: "completed")
5. Every task must be completed or deleted before final summary

Data discipline

• NEVER keep raw API responses or JSON in conversation context
• ALWAYS write intermediate data to temp files
• Subagent prompts MUST include: "Write results to . Report only counts and file paths."

4. Run enrichment

Dispatch a haiku agent to run the script:

Prompt: Run enrich-thin-frames.sh for fileKey <fileKey>.
Command: bash $CLAUDE_PLUGIN_ROOT/scripts/enrich-thin-frames.sh <fileKey>
Write stdout to /tmp/enrich-<fileKey>-result.txt.
Report only: total frames scanned, enriched count, skipped count, QMD status.

5. Report

Enrich complete: <fileKey>
  Frames scanned: N
  Enriched:       M  (were thin — re-fetched and frame.md regenerated)
  Already rich:   K  (skipped)
  QMD index:      updated | unchanged

Run /figma-differ:search "query" to verify enriched frames are now findable.