The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill processes external source code using read_file, creating a surface for indirect prompt injection. 1. Ingestion points: Untrusted data enters via read_file and grep_search on target files. 2. Boundary markers: Absent; there are no instructions for the agent to ignore commands embedded in the files. 3. Capability inventory: run_in_terminal and read_file. 4. Sanitization: Absent; the content is processed directly.
[Command Execution] (SAFE): The skill utilizes run_in_terminal to execute ruff and mypy. These are standard, reputable static analysis tools. In this context, the commands are static and do not incorporate unvalidated user-provided input into the command line execution, which minimizes risk.

code-reviewer