skills/simota/agent-skills/omen/Gen Agent Trust Hub

omen

Pass

Audited by Gen Agent Trust Hub on Oct 22, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting external data (specifications, design documents, and source code) and generating structured instructions ('Fix Prompts') for downstream agents like Builder or Beacon.
  • Ingestion points: Processes input from Scribe, Spark, and Magi, including design docs and feature proposals.
  • Boundary markers: Uses structured templates but lacks explicit safety delimiters to isolate processed content from its own instructions in the generated outputs.
  • Capability inventory: Generates prompts for Builder (code changes), Beacon (monitoring), and Triage (runbooks).
  • Sanitization: Not explicitly documented; relies on the agent's internal reasoning and template structure.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 22, 2026, 01:03 PM