ai-engineer
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to design and implement systems that ingest data from external, untrusted sources, introducing a potential surface for indirect prompt injection. 1. Ingestion points: The instructions reference web scraping, document processing, and API integrations as data sources (SKILL.md). 2. Boundary markers: The skill includes specific advice to treat retrieved text as untrusted and to implement prompt guardrails (SKILL.md). 3. Capability inventory: The skill outlines the use of tools for code execution, database queries, and web browsing (SKILL.md). 4. Sanitization: It recommends using moderation APIs, custom classifiers, and PII redaction strategies to mitigate risks.
- [NO_CODE]: The provided skill consists entirely of markdown-based instructions and metadata; no scripts, configuration files, or executable dependencies are included in the analyzed content.
Audit Metadata