The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data at several points, creating an attack surface where malicious data could influence agent behavior.
Ingestion points: Data is ingested through the read_output, list_skills, and get_run tools, which retrieve data from external skill executions and GitHub repository states.
Boundary markers: The instructions lack explicit boundary markers or delimiters when interpolating external outputs into the agent's context.
Capability inventory: The skill possesses significant capabilities, including run_skill, update_skill, update_strategy, update_soul, and install_pack (which creates pull requests).
Sanitization: No explicit sanitization or filtering or validation of the retrieved skill outputs is defined.
[SAFE]: The skill contains strong negative constraints preventing the agent from requesting API keys, tokens, or secrets, directing users to the vendor's official dashboard for credential management.
[SAFE]: All referenced domains (aeon.fun) are official resources belonging to the skill's author (aeonfun).

get-started