The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data in the form of movie scripts or fragments. This creates a surface for indirect prompt injection attacks where malicious instructions could be hidden within the script text to influence the agent's behavior. \n
Ingestion points: Processes full scripts or fragments provided by the user as specified in the '输入要求' (Input Requirements) section of SKILL.md. \n
Boundary markers: The instructions do not provide explicit delimiters or guidance for the agent to ignore potential instructions embedded within the user-provided scripts. \n
Capability inventory: The agent is configured with Read and Write tool permissions, which could be abused if an injection attack is successful. \n
Sanitization: There are no documented procedures for sanitizing or validating the input text before it is processed by the model.

script-evaluator