The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from an external API which could potentially contain malicious instructions. Ingestion points: The assessment and limitations fields returned by the compute-energy-shock endpoint in SKILL.md. Boundary markers: The Content safety section explicitly instructs the agent to treat responses as data only and disregard directive-like text. Capability inventory: The skill defines API interactions; it does not contain local scripts with subprocess or file-write capabilities. Sanitization: Relies on instructional guardrails to prevent the model from following commands found in external data.

assess-energy-shock