audit-code-health
Warn
Audited by Gen Agent Trust Hub on Mar 26, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run 'build' and 'tests' during the scan phase. This is a high-risk operation as it can lead to the execution of arbitrary, potentially malicious code defined in the build scripts or test files of an untrusted repository being audited.- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it processes untrusted source code without boundary markers or sanitization. 1. Ingestion points: target directory files scanned by the agent. 2. Boundary markers: absent. 3. Capability inventory: shell execution for build, tests, and audit commands. 4. Sanitization: absent.- [METADATA_POISONING]: The skill's metadata contains conflicting information; SKILL.md identifies the author as 'vercel', a trusted entity, while the repository URL and README point to 'kyzooghost'. This discrepancy could mislead users about the skill's origin and safety profile.
Audit Metadata