reality-check
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted user-provided content (plans, strategies, communications), which creates a surface for indirect prompt injection attacks.\n
- Ingestion points: The activation protocol in SKILL.md instructs the agent to "Parse the input" and "Assess completeness" of the work being reviewed.\n
- Boundary markers: The instructions do not specify any delimiters (like triple backticks or XML tags) or safety warnings to ignore instructions embedded within the reviewed content.\n
- Capability inventory: The skill utilizes web_search to research domain-specific information (referenced in SKILL.md under Core Process Flow and Domain Learning Protocol).\n
- Sanitization: There are no instructions for sanitizing or escaping the content before processing.\n- [SAFE]: No other security concerns were identified. The core instructions and reference files contain standard advisory frameworks and there is no evidence of obfuscation, remote code execution, or credential exposure.
Audit Metadata