The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were detected. The skill's behavior is consistent with its stated purpose of archiving project features.
[COMMAND_EXECUTION]: The skill utilizes git mv to move feature folders from .scratch/ to docs/archive/. This is a standard version control operation restricted to specific project subdirectories.
[DATA_EXPOSURE]: The skill reads local project files, specifically SPEC.md and ticket files in the .scratch/ directory, to verify completion status. It does not access sensitive system files (e.g., SSH keys, environment variables) or credentials.
[INDIRECT_PROMPT_INJECTION]: While the skill ingests data from user-controlled files (specifications and tickets), the processing logic is restricted to simple pattern matching for status lines (Status:) and checkboxes (- [x]). It does not execute commands found within these files, and the disable-model-invocation: true configuration prevents the model from being manipulated by potentially malicious text within the processed files during execution.

archive-feature