jev\_ampcode
Amp plugin that asks Jev to compare supplied alternatives against supplied evidence and priorities, as advisory input only.
Overview
Jev for Amp is a small V1 pilot plugin for the Amp coding agent that compares supplied alternatives against supplied evidence and user priorities. Jev returns typed choices and probabilities, not research, explanations or permission to act, and the README stresses it is not an Oracle replacement or a correctness gate.
Intended uses are comparing implementation approaches after inspecting code and constraints, comparing product or workflow options when priorities are known, and choosing between evidence-gathering actions that resolve a concrete uncertainty. It warns against using it for routine choices, reassurance, open-ended research, designing missing pieces, guessing what a user will authorize or proving safety, and advises one call per unchanged decision rather than repeating calls for endorsement. Only two runtime files are needed with no dependencies, running on Amp's Bun plugin runtime; the key is supplied as JEV_KEY via Amp's secrets UI, with TYPESAFE_API_KEY as fallback.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Given this evidence and these priorities, which supplied alternative is best? | — |
How it works
- Agent inspects code, constraints and priorities
- Agent supplies alternatives and evidence to the plugin
- Jev returns a choice with probabilities
- Agent uses the result as advisory input
Strengths
- Clear guidance on when not to use it
- No runtime dependencies
- Keys via Amp's native secrets
Caveats
- V1 pilot; advisory only
- Requires Amp with directory-plugin support
Getting started
Install index.ts and decision.ts as an Amp directory plugin and add JEV_KEY via Amp's secrets UI.At a glance
- Type
- Plugin
- Status
- prototype
- Language
- TypeScript
- Runtime
- Amp plugin (Bun)
- Has code
- yes
Use cases and integrations
How curated lists describe it
An Amp plugin for comparing supplied alternatives against supplied evidence and priorities.
logicrw/awesome-jev-projects
Frequently asked questions
What does the plugin decide?
Which of the alternatives the agent supplies best fits the supplied evidence and priorities.
Is it a correctness gate?
No. The README says it is advisory, not an Oracle replacement or a gate.
How is the key provided?
As JEV_KEY through Amp's secrets UI, or TYPESAFE_API_KEY as a fallback.
Related projects
- SkillRanker★ 134
Rust CLI that ranks which agent skills fit the next step from live session context with Jev, can abstain, and hooks into Claude Code.
CLIbetaRust Open-source Jev skills for Claude Code, Codex, Cursor and Grok: one client to gate, route, guard, verify, triage, review and act.
ProjectbetaTypeScriptLocal gateway and MCP tools that let Jev predict which tool a coding agent should call next, from suggestions to enforced choices.
MCP serverbetaTypeScriptFourteen opt-in Claude Code and Codex skills where Jev makes one bounded choice: which skill, model, test, file or next step.
ProjectbetaPythonToolkit where Jev condenses large MCP responses for Claude Code and maps Codex requests to reviewed, fixed repository operations.
ProjectbetaTypeScriptOmO/senpi agent plugin where Jev or Span-01 rates which skills and next tools fit the task and spots loops or completion.
Agent skillbetaTypeScript