jev-guard
By muratcakmak
Source description: Probability-scored guardrails for Claude Code: deny rule-breaking edits and unasked-for deploys, route your docs into each prompt, and check the final answer against the turn's own evidence.
Repository description
The repository describes its purpose as: Probability-scored guardrails for Claude Code: deny rule-breaking edits and unasked-for deploys, route your docs into each prompt, and check the final answer against the turn's own evidence.
Source evidence
The daily check found an explicit Jev / TypeSafe AI connection in the original README and implementation files at the same commit.
Scoring runs through [Jev](https://typesafe.ai) (`noul` questions — each returns a probability, not a yes/no), so the thresholds are yours to set. **Every failure path passes the call through**: no key, endpoint down, malformed response, the tool call proceeds. The guard never blocks on the scorer being unavailable.
Automatically source-checked: public repository, pinned README, and implementation files. Not run, benchmarked or security-audited.