loop-engineer

Pass

Audited by Gen Agent Trust Hub on Jul 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill facilitates automation by generating commands for system scheduling tools, such as crontab (Unix-like systems) and Windows Task Scheduler, to establish recurring execution triggers for user tasks. This allows the generated loop constructs to operate autonomously on a defined cadence.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it generates execution logic and agent instructions based on user-provided task descriptions without explicit sanitization or boundary isolation. \n
  • Ingestion points: User input provided during the diagnostic and interview phases in SKILL.md. \n
  • Boundary markers: The generated templates in references/artifact-templates.md do not include specific delimiters or safety instructions to isolate external content within the generated skills. \n
  • Capability inventory: The generated loop constructs can perform file operations in the ./loops/ directory and execute system commands to manage task schedules. \n
  • Sanitization: The skill does not perform sanitization or validation of user-supplied task definitions before using them to build new execution scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 22, 2026, 07:46 AM
Security Audit — agent-trust-hub — loop-engineer