cost-latency-optimizer
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill's implementation focus is on performance and cost metrics. It does not perform unauthorized network requests, access sensitive files, or attempt to persist across sessions.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input (prompts and queries) to perform optimizations. While it lacks explicit boundary markers to prevent the LLM from following instructions embedded within the data, the impact is limited to LLM invocation and hashed cache interaction.\n
- Ingestion points: The query and prompt strings processed in select_model_by_complexity and optimize_prompt in SKILL.md.\n
- Boundary markers: None present in the provided snippets.\n
- Capability inventory: Invocation of the llm() tool and set/get operations on a Redis cache.\n
- Sanitization: The skill uses hashlib.sha256 to generate cache keys, effectively preventing cache key injection attacks.
Audit Metadata