skills/wkentaro/skills/ask-me-again/Gen Agent Trust Hub

ask-me-again

Pass

Audited by Gen Agent Trust Hub on Aug 30, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill does not contain any code, external dependencies, or network operations. It is composed entirely of natural language instructions and configuration metadata. The 'disable-model-invocation: true' setting in the frontmatter is a security hardening measure that restricts the agent from invoking other models while this skill is active.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes content from previous assistant messages to generate interactive prompts. While this involves handling data from previous turns, the skill is limited to UI/UX reformatting and does not involve executing untrusted data or accessing sensitive resources.
  • Ingestion points: Preceding assistant messages as described in SKILL.md.
  • Boundary markers: Absent; the skill relies on the agent's interpretation of its own context window.
  • Capability inventory: Uses native interactive tools (AskUserQuestion, request_user_input) which are restricted to gathering user input.
  • Sanitization: None explicitly defined; the skill assumes the agent will handle text processing safely.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 30, 2026, 12:09 PM
Security Audit — agent-trust-hub — ask-me-again