write-template-playbooks
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied 'inventory rows' to generate instructions for building other agents, creating a path where untrusted input can influence agent configuration.
- Ingestion points: Inventory data (name, pitch, integrations, trigger) used as input in SKILL.md.
- Boundary markers: Templates lack explicit delimiters or instructions to isolate user-provided text from agent instructions.
- Capability inventory: The skill environment allows file manipulation (Write, Edit) and shell execution (Bash), alongside dynamic tool execution via test_run.
- Sanitization: The instructions focus on structural adherence and connection-based secret management rather than input content sanitization.
- [DYNAMIC_EXECUTION]: The skill facilitates the dynamic loading and execution of toolsets during the agent development process.
- Evidence: SKILL.md and references/worked-example.md describe using the test_run tool to exercise uncommitted tools via the delta.set.parameters.agent.tools parameter, allowing for runtime modification of the agent's capabilities.
Audit Metadata