incident

Pass

Audited by Gen Agent Trust Hub on Sep 1, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process and audit documentation files which are external data sources, creating a potential surface for indirect prompt injection attacks.
  • Ingestion points: The skill reads files from **/docs/runbooks/*.md, **/docs/incidents/*.md, and monitoring configurations.
  • Boundary markers: The agent is explicitly instructed to treat reviewed files as data, not instructions, and to identify/ignore common injection patterns.
  • Capability inventory: The skill is scoped to Glob, Read, and Write tools.
  • Sanitization: The agent is instructed to report injection attempts found in data rather than executing them.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 1, 2026, 06:48 AM
Security Audit — agent-trust-hub — incident