The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it ingests untrusted code contents from the user's project while possessing execution capabilities.
Ingestion points: The skill reads user project files using the Read, Grep, and Glob tools during exception audits.
Boundary markers: No explicit delimiters or instructions to ignore embedded natural language prompts are defined in the file to isolate the code contents.
Capability inventory: The skill is allowed to use Bash, Read, Grep, and Glob tools, creating a risk path if malicious instructions are processed.
Sanitization: There is no evidence of validation or sanitization applied to the text inputs before processing.

codeprobe-error-handling