The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using a local CLI tool (ao) to verify, record, and check the status of workflow gates (e.g., ao ratchet record).
[DATA_EXPOSURE]: The skill manages state by reading from and appending JSON data to a local project file located at .agents/ao/chain.jsonl. This activity is restricted to local workflow metadata.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by reading task status from an external file (.agents/ao/chain.jsonl).
Ingestion points: The skill reads the workflow history from .agents/ao/chain.jsonl using cat and tail in SKILL.md.
Boundary markers: None identified in the instructions for delimiting the processed JSON data.
Capability inventory: The skill has the capability to execute shell commands via the ao CLI and write to the local filesystem (SKILL.md).
Sanitization: There is no evidence of sanitization or validation of the JSON content before it is parsed or displayed by the agent.

ratchet