openrouter
Pass
Audited by Gen Agent Trust Hub on Mar 12, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists of 180 Markdown files providing documentation for the OpenRouter API. There are no executable scripts (.py, .js, .sh), binaries, or automation configurations included in the package.
- [NO_CODE]: As a documentation-only skill, there is no implementation logic that could be exploited. All code snippets provided in the documentation are for reference and use standard placeholders for credentials and identifiers.
- [PROMPT_INJECTION]: The content was scanned for patterns of prompt injection and behavioral overrides. No such instructions were found; the text is purely informative and uses natural instructional language describing API parameters and usage patterns.
- [DATA_EXFILTRATION]: There are no patterns suggesting unauthorized data access or exfiltration. API endpoints mentioned are official OpenRouter URLs, and examples for third-party integrations (e.g., Datadog, Sentry, Grafana) are documented neutrally as well-known observability services.
- [OBFUSCATION]: No signs of code or URL obfuscation were found. Placeholders for Base64 data and hex hashes are clearly identifiable as example data for multimodal API requests and authentication flows.
Audit Metadata