The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external documentation and existing codebase files which could contain adversarial instructions. Ingestion points: Local codebase files inspected during the workflow and external documentation or evidence retrieved via vendor tools. Boundary markers: The instructions lack specific delimiters or escaping rules for external content, though they mandate user confirmation for mutations. Capability inventory: Modification of local files and execution of vendor-defined capabilities. Sanitization: Instructions require the agent to list and inspect capabilities before use and verify all implementation results manually.
[NO_CODE]: The skill contains only instructional markdown and reference maps, with no executable scripts, binaries, or complex configuration files provided.

source-management