dev-swarm-stage-personas

Pass

Audited by Gen Agent Trust Hub on Mar 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is a purely instructional template for persona creation and does not contain code, network access, or privilege escalation.- [PROMPT_INJECTION]: The workflow involves reading user-provided research files, creating a surface for indirect prompt injection. This is mitigated by the scope of the skill's capabilities, which are limited to local document generation.
  • Ingestion points: ideas.md, 00-init-ideas/, 01-market-research/.
  • Boundary markers: None.
  • Capability inventory: Markdown file creation in the 02-personas/ directory.
  • Sanitization: None.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 13, 2026, 08:28 PM