hotpath
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates a workflow where the agent analyzes potentially untrusted external data, creating a surface for indirect prompt injection.\n
- Ingestion points: User performance complaints (SKILL.md) and diagnostic outputs such as ORM logs, APM traces, and DevTools flame charts (references/measurement.md).\n
- Boundary markers: The instructions do not define clear delimiters or specific instructions to treat external data as untrusted or to ignore instructions embedded within diagnostic reports.\n
- Capability inventory: The agent is directed to use system-level profilers, command-line benchmarking tools (e.g., k6, autocannon), and to modify application source code to implement optimizations.\n
- Sanitization: The skill lacks protocols for sanitizing, escaping, or validating the content of external logs or user input before processing.
Audit Metadata