dbs-knowledge

Pass

Audited by Gen Agent Trust Hub on Aug 25, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process local user data to build a knowledge base, which creates a potential surface for instructions embedded within user files to influence agent behavior.
  • Ingestion points: Processes all files within user-specified directories (SKILL.md).
  • Boundary markers: None explicitly mentioned to separate file content from agent instructions, though it focuses on navigational summaries (SKILL.md).
  • Capability inventory: Performs file system operations including reading, writing, moving, and listing directories (SKILL.md).
  • Sanitization: Includes a robust exclusion list that prevents the agent from reading sensitive files such as .env, .git, SSH keys, password databases, and large binaries (SKILL.md).
  • [DATA_EXPOSURE]: The skill performs broad scanning and reading of the user's local filesystem to create index files (SOURCE_OF_TRUTH.md).
  • Evidence: Scans directory structures up to multiple levels and reads file metadata and summaries (SKILL.md).
  • Mitigation: The skill instructions explicitly forbid reading sensitive browser data, private keys, or credentials, and limits its scope to a specific 'knowledge base' root directory (SKILL.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 25, 2026, 02:49 PM
Security Audit — agent-trust-hub — dbs-knowledge