write-plan
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external content such as tickets, PRDs, and specs to generate implementation plans. This creates a multi-step execution chain where malicious instructions in the source material could influence the plan and subsequently be executed by downstream agents (
implementorexecute-plan). - Ingestion points: The skill reads source asks from tickets, PRDs, specs, or chat conversations (referenced in §1 "Gather the inputs").
- Boundary markers: The template and instructions do not specify the use of delimiters or specific instructions to the LLM to ignore embedded commands within the source material.
- Capability inventory: While this specific skill is read-only (except for writing documentation), the plans it generates are designed to be consumed by agents with file-writing and command-execution capabilities.
- Sanitization: There is no evidence of sanitization or validation of the external input before it is interpolated into the planning process.
Audit Metadata