The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill analyzes user-provided project files (CSS, JSX, stylesheets) to suggest or apply transitions. This creates a potential surface where malicious instructions embedded in codebase comments or strings could attempt to influence the agent's behavior.
Ingestion points: The transitions review, transitions apply, and transitions refine commands scan the workspace for style declarations and component patterns in Component.tsx, .css files, and inline style blocks.
Boundary markers: The instructions do not define clear delimiters or "ignore" instructions for the content being analyzed, which increases the risk of the agent obeying instructions hidden in the code.
Capability inventory: The skill possesses the capability to suggest and apply code changes (file-write) across the project based on its analysis. It does not perform network operations or execute subprocesses.
Sanitization: No explicit sanitization or validation of the ingested code content is mentioned in the instructions.

transitions-dev