bullet-point-summary-first
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external, untrusted data such as articles and meeting transcripts.
- Ingestion points: The prompt templates in
SKILL.mddirect the agent to ingest external content through placeholders like[PASTE TEXT / DOCUMENT]and[PASTE TRANSCRIPT]. - Boundary markers: The provided templates do not include explicit delimiters or "ignore instructions" markers to isolate the ingested text from the core prompt logic.
- Capability inventory: The skill performs no subprocess calls, file writes, or network operations; its output is restricted to natural language text.
- Sanitization: There is no sanitization, escaping, or validation of the external content before it is processed by the agent.
- [NO_CODE]: The skill contains no executable scripts (.py, .js, .sh), binaries, or tool-calling logic. It is purely a collection of prompt engineering patterns and documentation.
Audit Metadata