recallmax
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill manages large volumes of external context, including RAG results and documents, which could contain hidden instructions intended to manipulate the agent's behavior.
- Ingestion points: Ingests 500K-1M tokens of external context, documents, and RAG results (SKILL.md).
- Boundary markers (present): Mentions source attribution and deduplication to prevent drift.
- Capability inventory: Auto-summarization, token compression, and fact-checking.
- Sanitization (present): Includes built-in cross-reference checks for fact verification.- [EXTERNAL_DOWNLOADS]: The skill references external repositories and a marketplace for installation and additional resources.
- References a GitHub repository at github.com/christopherlhammer11-ai/recallmax.
- References marketplace endpoints at genesis-node-api.vercel.app and genesis-marketplace.vercel.app, which are hosted on a well-known service provider.
- Provides an installation command npx skills add for adding the skill.
Audit Metadata