07-figma — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited 07-figma (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Construct screens in Figma via MCP following the design plan.
Parse $ARGUMENTS:
Hard dependencies:
docs/workflow-status.md — verify step 06 (Design Plan) status == completeddocs/Design/figma-design-plan.md exists with screen inventory + frame tree specsIf prerequisites fail:
/06-design-plan first."Pre-flight checks (beyond dependency clearance):
docs/Design/corrections-log.md (if exists). Note active corrections and graduated rules to apply during construction.docs/Design/figma-design-plan.md — full plan$ARGUMENTS specifies a batch or screen, scope to thatdocs/Design/{{PROJECT_NAME}}-design-document.md — tokens and component specs.impeccable.md — brand principles for design rationaleIf Figma file doesn't exist yet (Figma IDs table is empty):
figma-designer agent to:For each batch/screen in scope, delegate to figma-designer agent:
Construct [batch/screen] for {{PROJECT_NAME}} in Figma.
>
Design plan: [inline frame tree specs for this batch] Design tokens: [from design document] Consistency rules: [from design plan] Active corrections: [from corrections log] Graduated rules: [from design plan] Component reference: [node IDs from design plan]
>
Construction rules: - Batch 2-4 screens per MCP call (50KB JS capacity) - Use FILL sizing on children of auto-layout frames - Always use component instances (look up by node ID first) - Set padding on screen frame, not children - Return structured data (node IDs, names, dimensions) for each constructed element - Update design plan checklist on completion
The Figma agent reads many large files. To prevent context rot:
If construction fails or produces incorrect output:
"Construction failed/incorrect: [specific error or screenshot evidence]. Current state: [node IDs and observed properties]. Please fix: [specific instruction]."
blocked (reason) in docs/workflow-status.mddocs/workflow-status.md step 07 status to needs-review/correct. Run /07-figma [continue] to approve this batch and proceed to the next."[continue] after needs-review transitions to completed (or starts next batch)Step 07 — Figma Creation: [status]
Foundation: [complete/incomplete]
Screens constructed: [count]/[total]
Batch 1: [count]/[total] — [status]
Batch 2: [count]/[total] — [status]
...
Pre-flight:
Token sync: [OK / STALE — update needed]
Active corrections: [count]
MCP status: [connected / not tested]
Next actions:
- Review constructed screens in Figma
- Run /correct for any visual issues
- Run /07-figma [continue] for next batchdocs/workflow-status.md — extract step 07 sectiondocs/workflow-status.md step 07~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.