The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes several file system and version control commands to manage the project structure.
Evidence: Uses rm -rf, mv, and mkdir -p to manipulate directories within conductor/tracks/ as described in the 'Execute Archive', 'Execute Restore', and 'Execute Delete' sections of resources/implementation-playbook.md.
Evidence: Executes git add and git commit to track changes to the project state.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes data from files within the user's repository.
Ingestion points: Reads track information and status from conductor/tracks.md, metadata.json, and plan.md across multiple modes (List, Archive, Cleanup).
Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore potentially malicious content within these files.
Capability inventory: The skill has the ability to delete files (rm -rf), move files (mv), and commit changes to the repository.
Sanitization: While the skill mandates a specific ID format ({shortname}_{YYYYMMDD}) for new tracks, it does not specify sanitization or escaping for existing content read from the file system before it is used in logic or display.

conductor-manage