LLM Prompt Optimizer
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [No Code] (SAFE): The skill consists entirely of Markdown documentation and YAML metadata. There are no associated Python scripts, JavaScript files, shell scripts, or binaries. No commands are executed and no network requests are made.
- [Indirect Prompt Injection] (LOW): The skill defines workflows for the agent to ingest and process untrusted data (prompts provided by users for optimization). This creates an indirect prompt injection surface.
- Ingestion points: Workflow 1 ('Receive the current prompt') and Workflow 3 ('Identify target LLM').
- Boundary markers: The documentation explicitly recommends the use of delimiters (e.g., triple quotes or XML tags) as a best practice to mitigate accidental instruction obedience.
- Capability inventory: No dangerous capabilities are present. The skill lacks file-write access, network operations, or subprocess execution.
- Sanitization: No programmatic sanitization is defined; however, the skill provides instructional guidance on structural isolation of inputs.
Audit Metadata