Docs / source reviewed
jev-guardrails
Add Jev decision checks to a customer-support agent
How this project uses Jev
- Input
- Customer requests, agent responses, and proposed tool calls
- Jev decides
- Whether a request is in scope, safe, compliant with soft rules, or suited to a tool
- Code executes
- Code routes, reviews, or blocks actions and also runs deterministic output scanning
Evidence and limitations
The workflow and tools are simulated; reported evaluations were not reproduced. On Jev errors some checks return no decision and are skipped, so this is not an independent hard security boundary.
This project has not been run independently here. Author-reported results are not independently verified results.
Original sources
- https://github.com/deepansh-saxena/jev-guardrails/blob/475e70cbba9ee3ff674d4d9dba95681b2be05859/README.md ↗
- https://github.com/deepansh-saxena/jev-guardrails/blob/475e70cbba9ee3ff674d4d9dba95681b2be05859/tlife_agent/guardrails/backends/jev_backend.py ↗
- https://github.com/deepansh-saxena/jev-guardrails/blob/475e70cbba9ee3ff674d4d9dba95681b2be05859/tlife_agent/routines/base.py ↗
- https://x.com/Deepansh_AI/status/2102805949237125279 ↗