itut-recommendations
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides instructions for the agent to process, analyze, and format technical ITU-T Recommendations. This involves ingesting external document content, which represents an indirect prompt injection attack surface.
- Ingestion points: Draft Recommendations and contribution texts processed according to the structure defined in SKILL.md and references/rules-recommendation-structure.md.
- Boundary markers: The skill establishes a rigid clause structure (e.g., Clause 1: Scope, Clause 2: Normative references) that provides structural delimiters for ingested data.
- Capability inventory: The skill references the use of
officeclifor formatting operations (SKILL.md). - Sanitization: The instructions do not specify sanitization or escaping protocols for the processed external content.
- [EXTERNAL_DOWNLOADS]: The skill references an official Word template hosted by the International Telecommunication Union (itu.int). This is a well-known international organization, and the URL is provided for refreshing the local skeleton asset included with the skill.
Audit Metadata