parallel-docs-simplify-sync
Pass
Audited by Gen Agent Trust Hub on Mar 27, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill defines a prompt template for sub-tasks that interpolates user-provided 'Goal' and 'Scope' values. This creates an indirect prompt injection surface where adversarial content in the goal description could influence sub-agent behavior.
- Ingestion points: 'Goal' and 'Scope' parameters in the 'Task Prompt Template' within SKILL.md.
- Boundary markers: Uses markdown formatting for structure but lacks explicit instruction-isolation markers.
- Capability inventory: Sub-agents can modify project files (code, documentation, and skills) and run shell commands.
- Sanitization: Includes the constraint 'Do NOT touch files outside your track's scope' to limit sub-agent actions.
- [COMMAND_EXECUTION]: The completion checklist instructs the agent to run
pytestanduv run mkdocs build --strictto verify changes. These are standard local development tools used for testing and documentation building.
Audit Metadata