error-recoverer

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
  • [Indirect Prompt Injection] (LOW): The skill processes potentially untrusted error messages and execution context. * Ingestion points: handle_error function in SKILL.md. * Boundary markers: Not identified. * Capability inventory: System command execution and code modification. * Sanitization: None identified.
  • [Command Execution] (LOW): Recovery strategies include modifying file system permissions and paths which are high-privilege operations.
  • [Remote Code Execution] (LOW): The skill allows for automated dependency installation from external registries as a recovery method for environment errors.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:26 PM