The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and analyze untrusted source code, creating a potential vector for indirect prompt injection.
Ingestion points: The scan_file function within scripts/scan_cmdi.py reads the text content of files in a target directory to perform regex matching.
Boundary markers: There are no explicit boundary markers or instructions to ignore embedded prompts when the tool extracts and reports snippets of code to the agent.
Capability inventory: The skill is configured with access to Read, Bash(python3:*), Glob, and Grep tools.
Sanitization: The tool performs regex-based scanning and does not execute the ingested code; however, it does not sanitize or filter the content of findings which are then processed by the agent's context.

detecting-command-injection-patterns