The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's 'research' phase involves analyzing untrusted data from the user's project environment, which could contain adversarial instructions intended to redirect the agent's behavior.
Ingestion points: The skill reads project documentation (README), configuration manifests (package.json, pyproject.toml, *.tf, *.csproj), source code, and 'live UI' interfaces or screenshots.
Boundary markers: No specific boundary markers or 'ignore' instructions are used to separate the ingested project data from the agent's internal logic.
Capability inventory: The skill possesses the capability to read local files, access web content for UI analysis, generate and modify TypeScript code, and execute shell commands.
Sanitization: The skill instructions do not specify any validation or sanitization steps for the data processed during the research phase.
[COMMAND_EXECUTION]: The skill directs the agent to execute shell commands that are defined within the host project's development workflow.
Evidence: Instructions in SKILL.md require the agent to 'Run the app's own gates (lint, type-check, format, tests, build)' to verify the generated components.
[EXTERNAL_DOWNLOADS]: The skill depends on standard third-party libraries commonly used in the React ecosystem.
Evidence: The skill documentation identifies react (React 19) and motion (motion/react) as required dependencies for implementing animated scenes.

motion-scene