The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes source code from local solution files, which creates an attack surface where malicious instructions embedded in code (e.g., in comments or strings) could potentially influence the agent's behavior.
Ingestion points: The agent reads source code and IDE metadata from project files such as .cs, .cpp, and .xaml via IDE indexing tools as described in SKILL.md.
Boundary markers: The instructions do not specify explicit delimiters or "ignore instructions" markers when reading or processing the content of these code files.
Capability inventory: The skill possesses significant file-writing and code-modification capabilities through tools like rename_refactoring, move_type_to_namespace, and safe_delete (documented in the reference/tools/ directory).
Sanitization: No specific sanitization or filtering of the source code content is performed prior to processing it for refactoring actions.

refactoring-code