skill-fix-logs
Warn
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS] (MEDIUM): The skill instructs the agent to install 'repomix' from an external site if not present, which introduces unverifiable code into the environment.
- [COMMAND_EXECUTION] (MEDIUM): The use of 'repomix' and 'claude-code' involves executing system-level commands and modifying code based on potentially untrusted input.
- [DATA_EXFILTRATION] (MEDIUM): Targeted access to 'logs.txt' poses a risk of exposing sensitive data such as API keys or internal paths often found in logs.
- [PROMPT_INJECTION] (LOW): High surface area for Indirect Prompt Injection. The skill ingests untrusted data from logs and URLs without boundary markers or sanitization, and has the capability to execute commands. Evidence: 1. Ingestion: logs.txt, external URLs. 2. Boundary markers: Absent. 3. Capability inventory: repomix, claude-code. 4. Sanitization: Absent.
Audit Metadata