The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from multiple sources which could be used to influence agent behavior through embedded instructions.
Ingestion points: The skill reads After Effects project metadata via ae_project_info, animation states via ae_find_animation, and processes natural language feedback through the ae_review_motion tool's note parameter.
Boundary markers: There are no explicit instructions for the agent to use delimiters or ignore potential commands embedded within project metadata (like layer names or marker comments) or user-provided notes.
Capability inventory: The skill utilizes a powerful toolset for modifying the host environment, including ae_add_keyframes, ae_edit_keyframes, ae_apply_effect, and ae_layer_action. These allow for significant modification of the After Effects project.
Sanitization: While the skill suggests using ae_proof to validate the technical integrity of the motion (e.g., checking for strobing or timing drifts), it does not specify sanitization or filtering of input strings for instruction-like patterns.

ae-motion-principles