The Agent Skills Directory

[SAFE]: The skill files contain no malicious logic, command execution, or obfuscation. The behavior is consistent with the stated purpose of helping users follow through on goals.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied descriptions of goals and obstacles, which serves as an ingestion point for untrusted data. * Ingestion points: User query providing the 'Wish', 'Outcome', and 'Obstacle' inputs. * Boundary markers: The instructions specify using the template in references/TEMPLATE.md and explicit quality checks to constrain output. * Capability inventory: The skill has no dangerous capabilities such as file system writes, network requests, or shell command execution. * Sanitization: No input sanitization is present, but the lack of executable capabilities mitigates the risk of exploitation.

think-woop