work

Pass

Audited by Gen Agent Trust Hub on Mar 5, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted external data via the '${ARGUMENTS}' interpolation in the 'input_document' block, which introduces a surface for indirect prompt injection where a malicious document could attempt to override agent instructions.
  • Ingestion points: Untrusted work plans or specifications enter the agent context via the '${ARGUMENTS}' variable.
  • Boundary markers: The skill uses '<input_document>' XML-style tags to delimit the external content.
  • Capability inventory: The agent is granted access to high-impact tools including 'Bash' (for git and system operations), 'Write', and 'Edit' for filesystem modifications.
  • Sanitization: The 'Phase 1: Quick Start' section mandates that the agent read the plan, clarify ambiguities, and secure explicit user approval before performing any implementation tasks, serving as a critical manual verification control.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 5, 2026, 08:19 AM