The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script (log_results.py) located within the skill's own directory and uses the cat command to retrieve experiment status. These commands are used to aggregate and display results, which is the primary function of the skill.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from local .autoresearch JSON files and the output of its internal logging script to present a dashboard to the agent. While this creates a surface for data ingestion, the risk is negligible as it involves local experiment metadata.
Ingestion points: .autoresearch/{domain}/{name}/loop.json and output from {skill_path}/scripts/log_results.py.
Boundary markers: None present in the instructions.
Capability inventory: Execution of python and cat via shell.
Sanitization: Standard handling by the agent environment is assumed for the command arguments.

ar-status