moai-domain-humanize

Pass

Audited by Gen Agent Trust Hub on Jul 30, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill operates by ingesting and processing untrusted user-provided text for humanization, creating a surface for indirect prompt injection.
  • Ingestion points: External text content is accepted for processing as described in the multi-stage humanization workflow in SKILL.md and the various language modules.
  • Boundary markers: The skill requires the creation of an 'Invariant Ledger' (in SKILL.md) to define factual boundaries before editing and a 'Delta Audit' to verify survival of claims after editing.
  • Capability inventory: The skill utilizes standard file manipulation tools (Read, Write, Edit, Grep, Glob) as defined in the platform configuration in SKILL.md.
  • Sanitization: The instructions mandate strict preservation of factual anchors (numbers, dates, entities) and explicitly forbid inventing or modifying information during the humanization pass.
  • [NO_CODE]: The skill is implemented entirely through Markdown instructions and language-specific catalog modules. It does not include or install any executable code, scripts, or external software dependencies.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 30, 2026, 08:29 PM
Security Audit — agent-trust-hub — moai-domain-humanize