What premature-stop patterns does it actually intercept?

Empty turns, declared actions with no supporting tool calls, monitor promises with nothing running, and replies in the wrong language. The delivered nudge is a fixed stderr paragraph, not a generated critique.

How does it tell watching real work from an empty monitoring claim?

It compares the promise against a local count of background tasks, session crons, and unanswered cross-session handoffs. Zero running work blocks on the watch switch; mismatched work needs the stricter judge switch.

Why did I install it and nothing ever blocks a turn?

Install at user scope from the marketplace and set switches in settings.json, not shell exports, then restart. Installation alone does nothing; start with shadow mode, which calls Jev and logs without blocking.

What happens when the API key is missing or Jev times out?

Missing keys, timeouts, malformed bodies, out-of-range scores, and credential patterns in payloads all exit 0. That keeps a broken hook from permitting less work, at the cost of silently skipping the check.

How do I enable the watch and language checks?

Set STINGRAY_SHAPE3 and STINGRAY_LANG in the env block of ~/.claude/settings.json. Installation alone does nothing; the hook stays inactive until a switch is set, and you must restart Claude Code.

Can I try the checks without blocking turns?

Yes. With an API key, set STINGRAY_SHADOW=1, which calls Jev and logs decision records but never blocks turn completion.

Is stingray available for Codex as well as Claude Code?

Yes. It ships both a Claude Code plugin and a Codex plugin, installed from the GitHub marketplace, and CI checks that the two manifests carry the same version.

CommunityJev Agents & MCP

Stingray

Claude Code Stop hook that sends the final message to Jev to catch empty stops, unbacked action claims, unwatched background promises, and wrong-language replies.

See highlights

Open github.com

Stars: 6
Language: Shell
License: MIT
Published by: Community
Author: Nanako0129
Last verified: Oct 2026

Nanako0129/stingray

User repository on GitHub

View on GitHub

A Stop hook for Claude Code, Codex and Grok Build, for the turn that ends half-done — nothing done, an announced action never carried out, a promise to watch CI with nothing running — or in the wrong language. Every judgement is made by TypeSafe Jev.

Forks: 2
Last pushed: Updated Oct 2026

Repo stats from the GitHub API, cached Oct 2026.

About this project

Stingrays sit motionless on sand until stepped on. This hook operates the same way: silent during normal operation, intervening only when a turn stops prematurely. Every judgement goes to Jev: whether the turn did nothing, broke a promise, promised to watch something, or answered in the wrong language. The one thing counted locally is what is running in the background. Stop hook facts, measured not read.

Quoted from the project's README. Read the full README.

Tagscommunity awesomejev catalog-import shell

Works forJev for Classification Jev for Evaluation

Used inJev for Software & DevTools

Highlights

Judges no-action, broken-promise, unwatched-monitor, and wrong-language shapes via Jev.
Counts local background work exactly while treating the promise itself as a judgment.
Stays inert until switches are set, with shadow logging before enforcing blocks.
Fails open on every error so a broken hook never permits less work than baseline.

Watch out

Every failure path exits 0 and leaves the turn unexamined, which matches having no plugin rather than proving the turn was checked. Copy reviewed against the repository README on 2026-10-04, not from running it.

Questions about this entry

Short answers drawn from the linked project or page.

More Jev resources like this

pi-jev

Pi extension that pairs a local coding model with hosted Jev judgments for skill hints, file retrieval, and evidence ranking, with caching and failure bounds on every call.

Jev Agents & MCP#community#jevbest#catalog-import

Community

madeye/pi-jevTypeScript

jev-recipes

TypeScript recipe library with 248 focused functions that call Jev for routing, reranking, gating and labelling, returning structured ready-or-review decisions for Node backends.

Jev Agents & MCP#community#awesomejev#catalog-import

Community

agencyenterprise/jev-recipesTypeScript

hermes-jev-plugin

Hermes Agent plugin exposing Jev as four coding-workflow tools for parallel judgments, binary checks, option routing and rubric scoring, with a bundled question-design skill.

Jev Agents & MCP#community#jevbest#catalog-import

Community

ajensenwaud/hermes-jev-pluginPython

construct-auto-classifier

Shell gate for coding assistants that judges commands by effect and blast radius, asking Jev twelve questions or a chat model, with fast rules first.

Jev Agents & MCP#community#awesomejev#catalog-import

Community

godspede/construct-auto-classifierTypeScript

dsh-jev-prune

A DeepSeek Harness plugin that replaces DSH's size-based context compaction with Jev judgments: tool results that are still needed are never trimmed however large, and spent ones get a code-written receipt instead of a model-written summary.

Jev Agents & MCP#community#awesomejev#catalog-import

Community

yangyu666/dsh-jev-pruneJavaScript

clear-head

Claude Code Stop hook that checks each factual claim in a finished answer against the tool output read during that session, blocking the turn on claims the evidence contradicts or does not support.

Jev Agents & MCP#community#awesomejev#catalog-import

Community

VladyslavHontar/clear-headPython

Back to all resources

From the Jev community

Posts from builders shipping with Jev right now.

Follow @typesafeai