octocode-prompt-optimizer
Pass
Audited by Gen Agent Trust Hub on Aug 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is evaluated as safe. It is a well-structured instructional document intended to help agents self-optimize their instructions based on prompt engineering best practices.
- [PROMPT_INJECTION]: While the skill contains strong directive language such as 'CRITICAL
- FORBIDDEN' and 'Triple Lock' patterns, these are internal self-regulation mechanisms designed to enforce quality and prevent logic drift during the optimization process. They do not target the bypass of external AI safety filters.
- [DATA_EXFILTRATION]: No network tools or data exfiltration patterns were detected. The skill allows 'Read-only file access' and 'Safe file edit/write' strictly for processing prompt files as requested by the user, and these tools are mapped to local runtime aliases.
- [COMMAND_EXECUTION]: The skill restricts tool usage to specific file-related operations necessary for its primary purpose. It explicitly forbids the execution of commands unrelated to prompt optimization.
- [PROMPT_INJECTION]: The skill processes untrusted input (the prompts to be optimized), which is a surface for indirect prompt injection. However, the skill implements a rigorous 6-step gated workflow (Read, Understand, Rate, Fix, Validate, Output) with mandatory validation checks that require the agent to verify intent preservation, mitigating the risk of instructions within the processed text hijacking the agent's behavior.
Audit Metadata