The Agent Skills Directory

[SAFE]: No security vulnerabilities or malicious patterns were identified. The skill follows best practices for managing project documentation and local file organization.
[DATA_EXPOSURE]: The skill processes project-related data such as business rules, user research, and API designs. All operations are confined to the local filesystem (moving files from temp/ to SoT/ or archive/), with no indications of data exfiltration or unauthorized access to sensitive system files.
[COMMAND_EXECUTION]: The skill instructions mention the use of a tool named ghm-id-register for generating document IDs. This appears to be a local project-specific utility. The skill uses standard filesystem tools (Read, Write, Edit, Glob, Grep) as permitted in its configuration for its intended purpose of document management.
[PROMPT_INJECTION]: While the skill ingests untrusted content from temporary files (indirect prompt injection surface), this is inherent to its primary function as a documentation processor. The skill provides structured templates and clear logical boundaries for the extraction process, which mitigates accidental execution of instructions embedded in the processed data.

ghm-harvest