The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted source code from external repositories which could contain malicious instructions.
Ingestion points: Project files (Python, JS, TS, Rust, YAML, etc.) analyzed during static analysis, security scans, and code smell detection.
Boundary markers: None; there are no instructions or delimiters provided to the agent to distinguish between the content of the analyzed code and instructions.
Capability inventory: Execution of local static analysis tools and shell commands (uv, bunx, cargo, semgrep, rg).
Sanitization: None; the skill reads raw file content and reports findings directly to the agent's context without escaping or filtering.
[EXTERNAL_DOWNLOADS]: The skill uses well-known development tools that may be downloaded or updated from official registries during execution.
Evidence: The instruction semgrep --config=auto . (install if missing) suggests a potential download of the Semgrep utility.
Evidence: The command bunx jscpd executes a package from the npm registry.
Context: These are standard industry tools for static analysis and are used as intended for a code auditing skill.

analyze-code