The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface where untrusted data could influence agent actions.
Ingestion points: The skill reads source facts and transcript timings using the get_media_context tool (SKILL.md).
Boundary markers: No delimiters or "ignore instructions" warnings are defined for the data retrieved from external media sources.
Capability inventory: The agent has access to sensitive tools including create_brief, compose, create_render, and get_account for resource management and network operations (SKILL.md).
Sanitization: There is no evidence of sanitization or validation for the content processed during the beat planning phase.
[COMMAND_EXECUTION]: The skill interacts with system-level binaries and custom tools for video composition.
Evidence: The instructions reference cueframe-cli and a binary named cueframe for managing project mechanics (SKILL.md).
Evidence: The skill utilizes a suite of composition operations such as create_render and score_composition to execute the planned video flow (SKILL.md).

cueframe-storyboard