The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted data from external translation files, creating a surface for indirect instructions to reach the agent.
Ingestion points: scripts/loc-sheet.js reads translation content from translated.csv and glossary.csv provided by the user.
Boundary markers: There are no explicit delimiters or safety instructions wrapping the imported text content.
Capability inventory: The skill has access to extensive After Effects automation tools via ae_run_jsx, ae_batch, and ae_render_frame as seen in SKILL.md and the provided scripts.
Sanitization: No sanitization or filtering is performed on the translated text before it is used to generate command payloads.
[DYNAMIC_EXECUTION]: The skill utilizes dynamic script evaluation and After Effects expressions to perform its tasks.
Evidence: In SKILL.md, the ae_run_jsx tool is invoked using $.evalFile to execute the local scripts/loc-duplicate.jsx script.
Evidence: The documentation describes implementing After Effects expressions that dynamically load and parse string data from strings.json at runtime.

ae-localization