The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process and analyze untrusted project source code, which constitutes a potential injection surface.
Ingestion points: Reads src/index.ts and other imported modules in the project using Read and Grep.
Boundary markers: The instructions do not define delimiters or warnings to ignore instructions that might be embedded in the analyzed code.
Capability inventory: The skill uses Read, Glob, Grep, and Bash to interact with the environment.
Sanitization: The agent processes raw source code strings without specific sanitization or escaping mechanisms.

sync-validate