implement-automation-checklist

Pass

Audited by Gen Agent Trust Hub on May 13, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by processing data from an external source (the design scheme from assess-automation-checklist) to influence code generation and verification tasks.\n
  • Ingestion points: SKILL.md (Step 1: "读取 assess-automation-checklist 输出的设计方案").\n
  • Boundary markers: Absent; the instructions do not define delimiters or markers to isolate the external content from the agent's internal logic.\n
  • Capability inventory: The skill involves writing executable TypeScript files to the filesystem and running them for integration testing.\n
  • Sanitization: Absent; no validation or sanitization of the input design scheme is mentioned.\n- [COMMAND_EXECUTION]: The skill facilitates the creation and execution of code within the local environment as part of its primary function.\n
  • Evidence: workflow.md describes the generation of an executable checker/ project structure, and checklist.md instructs the agent to verify that the script is "runnable in target project."
Audit Metadata
Risk Level
SAFE
Analyzed
May 13, 2026, 11:55 AM
Security Audit — agent-trust-hub — implement-automation-checklist