dbs
Pass
Audited by Gen Agent Trust Hub on Jul 27, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting and acting upon the output of previous skills in its navigation mode.
- Ingestion points: Analyzes the 'specific conclusion or key signals' from the output of preceding skills (e.g., /dbs-diagnosis, /dbs-benchmark) in SKILL.md.
- Boundary markers: The instructions lack explicit delimiters or safety warnings to distinguish between control instructions and data content from the ingested skill outputs.
- Capability inventory: The skill primarily performs routing and navigation, but facilitates the execution of other skills that may have write or network capabilities (e.g., /dbs-save, /dbs-agent-migration).
- Sanitization: No evidence of sanitization or validation of the ingested content before it is used to influence the agent's next action.
- [PERSISTENCE_MECHANISMS]: The skill references functionality for storing and retrieving decision logs and session states in the local filesystem at '~/.dbs/decisions/'. This is intended for long-term project tracking and project restoration.
Audit Metadata