The Agent Skills Directory

[SAFE]: The skill incorporates a robust 'Sensitive Information Prohibition Rule' that mandates the screening and masking of PII, credentials, and internal domain details. It provides the agent with specific regex-based shell commands to verify content safety before any documentation is stored.
[COMMAND_EXECUTION]: Employs local shell commands (grep, mkdir, ls) to detect environment paths and perform safety validations on generated feedback content. These operations are restricted to the local filesystem and do not involve unauthorized network communication.
[SAFE]: The operational scope is strictly limited to generating feedback documentation in a designated directory. The instructions explicitly forbid the modification of the target skill's source code, reference files, or configuration.

peach-skill-feedback