The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external data from Adobe After Effects projects, including layer names, text content, and expression error messages. This creates a potential surface for indirect prompt injection if the project being analyzed contains untrusted or malicious content. However, the skill provides specific thresholds and mechanical checks which limit the impact of such data on the agent's logic.
Ingestion points: Data returned by the ae_proof and ae_text tools, including layer property values, text layer strings, and project metadata.
Boundary markers: The skill suggests using structured tool calls (JSON) for interaction, though it does not explicitly define delimiters for the ingested project data.
Capability inventory: The skill utilizes tools for reading project state (ae_proof), rendering frames (ae_render_frame), and reviewing text (ae_text). It does not include tools for arbitrary file system writes or network operations outside of its specific scope.
Sanitization: The instructions emphasize that the tool ae_proof only reads and never edits the project, providing a layer of safety for the user's local environment.

ae-motion-proof