synabun-memory
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill operates on a persistent memory store, creating an inherent surface for indirect prompt injection where historical data could contain malicious instructions. The skill proactively addresses this by including explicit safety markers: SKILL.md states that recalled text cannot override current instructions or permissions, and QA.md reinforces that recalled content must remain untrusted evidence. Ingestion occurs via the recall and memories tools, while the skill maintains capabilities to write, update, and delete records as part of its primary function.
- [EXTERNAL_DOWNLOADS]: The skill references external technical resources to ensure compatibility with the Synabun toolset. These references are documented in references/tool-calls.md and point to documentation at synabun.ai and specific source code blobs in a GitHub repository for schema verification. These are informational references and do not involve the execution of remote code or the installation of untrusted dependencies.
Audit Metadata