The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection via the files it analyzes.
Ingestion points: File system metadata (directory names, file names) and project manifest files like package.json, Cargo.toml, or pyproject.toml are read during execution.
Boundary markers: Absent. There are no instructions to the agent to ignore or delimit embedded commands found within the analyzed file structures.
Capability inventory: The skill uses the Bash tool to execute commands like find, wc, and tree, which could be influenced by malicious content in the codebase being mapped.
Sanitization: No sanitization or validation is applied to the data retrieved from the file system or manifest files.

file-analysis