The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes shell commands including rm -rf, mv, mkdir, and git to manipulate the file system within the conductor/ directory. While these operations are necessary for lifecycle management, the use of rm -rf on paths derived from identifiers requires caution. The skill effectively mitigates this risk by requiring the exact string 'DELETE' for confirmation and verifying that the target directory exists before execution.- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes content from local files that could be influenced by external sources.
Ingestion points: conductor/tracks/{trackId}/metadata.json, conductor/tracks/{trackId}/plan.md, and conductor/tracks.md.
Boundary markers: The skill does not define explicit delimiters or instructions for the agent to ignore potentially malicious embedded commands when reading these files.
Capability inventory: The skill has the capability to delete files (rm -rf), move directories (mv), and create git commits (git commit).
Sanitization: No explicit sanitization or validation of the file content is performed beyond checking for file existence.

conductor-manage