allium-voice

Pass

Audited by Gen Agent Trust Hub on Sep 1, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill contains natural language instructions providing stylistic guidance for writing software specifications. It does not include any code, external scripts, or network operations.- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process and edit user-supplied .allium files. Although this involves processing untrusted data, the instructions are limited to stylistic guidance and do not involve high-risk capabilities.
  • Ingestion points: .allium files, review comments, and KDoc as specified in the frontmatter and prose sections.
  • Boundary markers: No explicit delimiters or boundary markers are defined within the skill instructions.
  • Capability inventory: The skill is restricted to text editing and formatting within the context of the agent's file editing tools; no subprocess, network, or eval capabilities are invoked.
  • Sanitization: No specific sanitization or filtering of external content is mentioned.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 1, 2026, 07:24 PM
Security Audit — agent-trust-hub — allium-voice