arkweb-brainstorm

Pass

Audited by Gen Agent Trust Hub on Aug 8, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection Surface. The skill interpolates user-provided requirement descriptions and external evidence into tasks for specialized subagents.
  • Ingestion points: Data is ingested through the '{需求描述}', '{user_requirement}', and '{deepwiki_hits}' variables in SKILL.md.
  • Boundary markers: The skill uses markdown headers like '## 用户需求' to delimit input but does not include explicit instructions for the subagents to ignore potential commands embedded within that content.
  • Capability inventory: The skill is capable of spawning up to 10 subagents and performing file read and write operations on the local project directory.
  • Sanitization: No input validation or escaping logic is applied to the processed requirements or external data hits.
  • [COMMAND_EXECUTION]: Subagent Orchestration. The skill manages a workflow by spawning subagents located in local vendor-controlled paths ('{SKILL_HOME}/skills/arkweb-experts/'). This behavior is standard for the skill's purpose as an architectural coordinator.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 8, 2026, 09:00 AM
Security Audit — agent-trust-hub — arkweb-brainstorm