legacy-code-summarizer
Pass
Audited by Gen Agent Trust Hub on Mar 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a wide range of shell commands to explore the project structure and search for patterns, including 'find', 'grep', 'wc', and 'tree'.\n- [REMOTE_CODE_EXECUTION]: The workflow involves running the target project's test suite using tools like 'pytest', 'mvn clean test', or 'npm test'. If the codebase being analyzed is malicious, this could result in the execution of arbitrary code within the agent's environment.\n- [EXTERNAL_DOWNLOADS]: The skill suggests installing several well-known code analysis and auditing tools from official registries during execution, such as 'pylint', 'radon', 'bandit', 'pip-audit', and 'madge'.\n- [PROMPT_INJECTION]: The skill processes untrusted external code to generate summaries, making it susceptible to indirect prompt injection.\n
- Ingestion points: The agent reads source files, configuration manifests, and documentation throughout the analyzed project.\n
- Boundary markers: There are no instructions or delimiters provided to ensure the agent ignores potential commands or misleading instructions embedded within the codebase.\n
- Capability inventory: The agent possesses subprocess execution, file system access, and network reachability for package management.\n
- Sanitization: Content from the analyzed files is not sanitized or escaped before being interpreted by the LLM.
Audit Metadata