dbs-knowledge
Pass
Audited by Gen Agent Trust Hub on Jul 18, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill operates by reading and processing local user files, which creates an attack surface for indirect prompt injection if those files contain malicious instructions.
- Ingestion points: Local directories specified by the user as knowledge base roots.
- Boundary markers: The skill uses
SOURCE_OF_TRUTH.mdas an index, but does not specify delimiters for the content of individual files processed by the agent. - Capability inventory: The skill allows the agent to read, write, move, and copy files locally.
- Sanitization: No explicit sanitization or filtering of file content is described beyond excluding sensitive file types.
- [COMMAND_EXECUTION]: The skill instructs the agent to perform file system operations including directory creation, file moving, and file writing to maintain the knowledge base structure.
- Scope: Operations are restricted to the user-defined knowledge base directory.
- Safeguards: Explicit instructions require user confirmation before any batch movement, renaming, or deletion of files.
Audit Metadata