The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the agent-team CLI tool to perform maintenance tasks such as checking, updating, and cleaning skill artifacts. These actions are limited to cache management as defined in the skill boundaries.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from local configuration files and the skill cache, which creates an attack surface for data-driven instructions.
Ingestion points: Reads from .agent-team/rules/index.md and the skill cache.
Boundary markers: None identified.
Capability inventory: CLI command execution via agent-team for maintenance.
Sanitization: No specific validation or escaping of ingested data is detailed.

skill-maintenance