skills/fradser/dotclaude/create-prd/Gen Agent Trust Hub

create-prd

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes a Bash tool to interact with the lark-cli utility for creating and updating documents on the Feishu/Lark platform. The execution is strictly scoped to the lark-cli binary in the platform configuration, mitigating the risk of arbitrary shell command execution. PRD content is passed as a string argument to this tool.
  • [INDIRECT_PROMPT_INJECTION]: The skill implements functionality to read existing project documents from docs/ or prd/ directories to bootstrap the PRD creation process. This ingestion of external data is mediated by a manual confirmation step in Phase 2, where the agent is instructed to present all pre-filled information to the user for verification, which serves as a control against instructions potentially embedded in the source documents.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 11:53 PM