by shiftynick · Codex Skill · ★ 41
jev-axi for TypeSafe's Jev model, for the Agent eXperience Interface conventions it follows. An AXI (agent-ergonomic CLI) for TypeSafe's Jev, a System One model that answers typed questions about text with calibrated probabilities in about half a second. It never generates text; it makes judgments. Coding agents use it where a fast, cheap, calibrated call beats reading or reasoning: Safety: block risky agent tool calls before they run (), gate shell commands in scripts and cron jobs (), stop commits that add credentials (), and screen fetched pages, issues, and vendored docs for prompt...
| Stars | 41 |
| Forks | 9 |
| Language | TypeScript |
| Category | Codex Skill |
| License | MIT |
| Quality Score | 70.8555013689712/100 |
| Open Issues | 8 |
| Last Updated | 2026-10-15 |
| Created | 2026-09-16 |
| Platforms | claude-code, cli, codex, node |
| Est. Tokens | ~18k |
These tools work well together with jev-axi for enhanced workflows:
Explore other popular codex skill tools:
jev-axi is Agent-ergonomic CLI for TypeSafe's Jev: fast calibrated judgments (pick, rate, check, rank, triage, guard) from the shell. It is categorized as a Codex Skill with 41 GitHub stars.
jev-axi is primarily written in TypeScript. It covers topics such as agent-safety, ai-agents, axi.
You can find installation instructions and usage details in the jev-axi GitHub repository at github.com/shiftynick/jev-axi. The project has 41 stars and 9 forks, indicating an active community.
jev-axi is released under the MIT license, making it free to use and modify according to the license terms.