pencil-design-bridge

Pass

Audited by Gen Agent Trust Hub on Jul 29, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [SAFE]: The skill includes explicit instructions to prevent the storage of sensitive data such as API keys, tokens, or private credentials in design handoff files.
  • [COMMAND_EXECUTION]: The SKILL.md file documents a validation process involving the execution of local Python scripts for linting and schema verification, using environment variables to locate the tools.
  • [SAFE]: Access to .pen design files is restricted to specific MCP tools, explicitly forbidding the use of standard shell commands like cat or grep to ensure artifact integrity.
  • [PROMPT_INJECTION]: The skill ingests design briefs, representing an indirect prompt injection surface. Ingestion points include the 'brief' and 'review-output' modes in SKILL.md. Boundary markers are established via the references/pencil-bridge-contract.md and the mandatory confirmation of approval by the senior-ui-ux-orchestrator agent. Capability inventory is limited to design tool interactions and local validation scripts. Sanitization is managed through manual workflow checkpoints that verify source authority.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 29, 2026, 07:31 AM
Security Audit — agent-trust-hub — pencil-design-bridge