Agent tooling — page 45

Session-pinned model and thinking selection for Pi using TypeSafe Jev.

Agent memory that decides what to hydrate — local topology recall + TypeSafe Jev gates (not another dump-everything store).

SEAL — No seal, no advance. Workflow kernel: product truth as a sealed graph (not an agent parade, not Jev-with-UI).

Deterministic pre-agent skill routing for Hermes agents: audit, canonical catalog, profile policy, and bounded Jev judgment.

Portable skills for coding agents: review changes, verify completion claims, and make typed decisions with Jev, local models, or your own server.

Guardrails for AI agent actions. Deterministic policy, Jev semantic checks, and replayable decisions.

Cache-aware, TypeSafe-driven model tier router for Claude Code

Fail-closed Jev policy gate for Hermes Agent tool calls

Open WebUI pipeline that routes prompts to a cheap or expensive LLM using TypeSafe AI classification

Experimental out-of-tree Kubernetes scheduler plugin that uses TypeSafe's Jev model to produce typed, advisory diagnoses for unschedulable Pods.

See what an AI agent actually did, not just its score. Labels every step of a CLI agent's run and compares two rounds against the noise a re-run produces anyway.

A stop-loss for coding agents. Counts loops in code, judges drift with TypeSafe's Jev, and never sends your source anywhere.

Intercepts destructive shell commands before they run. ~2ms local fast-path, context-aware risk scoring via TypeSafe Jev, deterministic offline fallback. Single Rust binary for zsh, bash and fish.

Typed state, explicit transitions, one narrow model judgment, and a human approval gate — a rebuildable agent workflow using and testing Jev.

Bare-bones OpenRouter + TypeSafe Jev classifier: route legal prompts to Texas statute databases

TypeSafe Jev (System One decision model) as a native jev_decide tool plugin for DeepSeek Harness

GitHub Action for agent-written PRs: one comment that says what the PR really does, which hunks need a human, why, and what to verify. About a cent per PR, on TypeSafe Jev.

TypeSafe/Jev decision signals for Hermes: turn classification, conservative tool shaping, approvals, and coding verification.

An independent, evidence-first field guide to Jev (TypeSafe AI's System One model) — for people and for coding agents. Not affiliated with TypeSafe AI.

Score every Cognigy conversation against rubrics you write, using TypeSafe Jev instead of an LLM. CLI, local web UI, and an agent skill.

Context compaction for agent harnesses: a decision model scores transcript spans and tombstones the ones that stop mattering. Includes an MCP restore server.

把决策模型（typesafe/jev-1.13，经 OpenRouter 的 decisions 端点调用）封装成业务决策工具：注册表驱动，带置信度门控与硬约束。非官方项目。

A liquid magic eight ball powered by TypeSafe Jev decisions through OpenRouter

Harness for TypeSafe's Jev decision model: typed questions over datasets via OpenRouter, JSONL answers, summary reports.