The Agent Skills Directory

[SAFE]: The skill facilitates state recovery by directing the agent to read local configuration and context files (worker.yaml, task.yaml, context.md). These actions are restricted to the agent's workspace and are necessary for functional continuity.
[COMMAND_EXECUTION]: The skill specifies the use of the agent-team task show CLI command to refresh task details. This is a legitimate use of a project-specific tool for task management.
[INDIRECT_PROMPT_INJECTION]: The skill defines a surface for data ingestion by instructing the agent to process content from worker.yaml, task.yaml, and context.md.
Ingestion points: worker.yaml, task.yaml, and context.md (SKILL.md).
Boundary markers: None specified.
Capability inventory: Local file reading and execution of the agent-team task show CLI tool.
Sanitization: Not specified.

worker-recovery