executing-plans

Pass

Audited by Gen Agent Trust Hub on Jun 23, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill operates by reading and executing instructions from an external source (an implementation plan), which constitutes an indirect prompt injection attack surface.
  • Ingestion points: The skill loads and processes external implementation plan files in Step 1 (SKILL.md).
  • Boundary markers: Absent. There are no explicit instructions for the agent to use delimiters or to treat the content of the plan files as untrusted data.
  • Capability inventory: The skill references capabilities for writing code, running verifications, and managing git workspaces through integrated sub-skills (SKILL.md).
  • Sanitization: Absent. The skill does not specify any validation or filtering of the content found within the plan files before execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 23, 2026, 01:31 PM
Security Audit — agent-trust-hub — executing-plans