aesthetic-expansion-kanban

Pass

Audited by Gen Agent Trust Hub on Jul 10, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it ingests untrusted data from external sources and local project plans. 1. Ingestion points: User scope, project plans, and research corpora generated from external websites (SKILL.md). 2. Boundary markers: Research data is partitioned into specific directories, though explicit delimiters for external text are not mandated. 3. Capability inventory: The skill performs file system writes and executes shell commands for testing (pnpm) and Kanban management. 4. Sanitization: Content is processed through a visual analysis framework and review gates before implementation.
  • [COMMAND_EXECUTION]: The skill involves running standard project tools and Kanban management commands limited to the project's worktree for software development tasks.
  • [EXTERNAL_DOWNLOADS]: The workflow incorporates data collection from external URLs for research purposes, with explicit instructions to prioritize public access and record limitations.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 10, 2026, 08:04 AM
Security Audit — agent-trust-hub — aesthetic-expansion-kanban