code-check
Pass
Audited by Gen Agent Trust Hub on Aug 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a wide array of local development tools including compilers (tsc, go build), linters (ESLint, flake8, clippy), and test runners (jest, pytest, cargo test). While these are standard tools, executing them involves running code defined within the project repository.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8). It ingests untrusted project source code and processes the output of tools that analyze this code.
- Ingestion points: Project source files (JS, TS, Python, Rust, Go, Swift) are read by tools and the agent to identify and fix errors.
- Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore embedded instructions within the code or tool outputs.
- Capability inventory: The agent has the capability to execute shell commands and write modifications to the local filesystem.
- Sanitization: No sanitization or validation of the content being 'fixed' is specified, which could allow malicious instructions in code comments or test descriptions to influence agent behavior.
Audit Metadata