The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes Git commands to manage project history. It uses git log to find SHAs and git revert or git reset to undo changes. While git reset --hard is destructive, the skill requires explicit user confirmation and provides a warning about potential data loss.
[INDIRECT_PROMPT_INJECTION]: The skill processes task descriptions from local files which could be manipulated to influence agent behavior.
Ingestion points: Reads from conductor/index.md, conductor/tracks.md, and implementation plan files (e.g., conductor/tracks/<track_id>/plan.md).
Boundary markers: The skill relies on markdown parsing for tracking status ([~], [x]) but does not implement explicit delimiters or instructions to ignore embedded prompts within task descriptions.
Capability inventory: Performs Git reverts, hard resets, and file modifications via terminal commands and file system tools.
Sanitization: No explicit sanitization or filtering of task content is performed before it is presented to the user as an option or used in the revert logic.

conductor-revert