Jev directory
CommunityJev Agents & MCP
Dsh Auto Mode
A DeepSeek Harness auto mode hosted on GitLab: full-access automation guided by TypeSafe System One at every end prompt, MIT-licensed and created in September 2026.
- Reading time
- 2 min
- Published by
- Community
- Author
- allenh1
- Last verified
- Oct 2026
- Capabilities
- Open source
Highlights
- Hosted at git.allen-software.com, not GitHub, so tooling that assumes github.com needs adjusting.
- DeepSeek Harness Auto mode offers full access guided by TypeSafe System One at every end prompt.
- Project tags read jev, typesafe, system-one, deepseek-harness, cordis, and ai-agent.
- MIT-licensed, created September 17, 2026, with 15 commits on a single branch and no tags.
Watch out
Copy reviewed against the GitLab project page on 2026-10-15; the page is a thin listing whose README section renders as 'Loading', so details beyond the description, licence, tags, and commit counts were not reviewable.
Questions about this entry
Short answers drawn from the linked project or page.
A mode for the DeepSeek Harness that runs with full access guided by TypeSafe System One at every end prompt, per the project's own one-line description on its GitLab project page.
On a self-hosted GitLab at git.allen-software.com under the allenh1 account, rather than GitHub, which is also why the yibie list links the GitLab address directly.
The description pairs full access with System One guidance at each end prompt, so the harness checks decisions with Jev at the point where the agent would otherwise stop or hand back control, instead of relying on manual approval rules.
MIT, per the project page's licence badge. The project was created on September 17, 2026, and shows no releases or tags yet.
Modestly: the page showed 15 commits on one branch and no tags as of 2026-10-15, so it reads as an early single-author project rather than a maintained library.
It lists TypeSafe System One in its own description and tags the project jev, typesafe, and system-one, meaning the harness consults Jev's typed decisions rather than free-form model output at each end prompt.
It appears in yibie/awesome-jev's listing of community harness experiments, which is how it entered this catalog; the project page itself does not link back to any list.
More Jev resources like this
A small, opinionated harness setup for the Pi coding agent: scope control keeps tasks to declared files, an improvement loop turns friction into tested changes, eight pinned packages are verified before install, and LANCET Nano screens shell commands.
Jev Agents & MCP#community#catalog-import#logicrw
Community
tannermidd/specpiJavaScript
tannermidd/specpiJavaScriptAn intent compiler for AI agents: it probes what files and tickets already answer, asks only what they can't, and emits a typed IntentSpec YAML saved to .intent/<intent>.intent.yaml before work is routed. The README positions it as the input layer for routers and typed-decision models like Jev and Laya.
Jev Agents & MCP#community#catalog-import#madewithjev
Community
angel291592/intent-routerPython
angel291592/intent-routerPythonA developer-preview personal agent harness that turns a sequence of model calls into continuing work: understand the request, choose an action, execute, inspect evidence and recover context. You pick the general model and can optionally connect a Reflex/Jev decision service.
Jev Agents & MCP#community#github#agent
A guard for coding agents that puts Jev in front of every step: it judges tool calls for intent and risk, scans tool outputs for instructions aimed at AI agents, and checks replies, mapping probabilities to allow, ask or warn.
Jev Agents & MCP#community#catalog-import#logicrw
Community
harshwasan/pi-jev-sentinelTypeScript
harshwasan/pi-jev-sentinelTypeScriptPrompt reviewer for Claude Code and Codex CLI: Jev first decides whether a submitted prompt needs a review, and when it does, a lightweight model flags overlooked context, unstated assumptions, or the real problem in a local web UI.
Jev Agents & MCP#community#catalog-import#logicrw
Community
jsk4581/jev-blindspotTypeScript
jsk4581/jev-blindspotTypeScriptClaude Code plugin that asks Jev a few typed questions before every turn and sets reasoning effort, subagent model and effort, strategy, and the skill a prompt needs; you keep Opus for the conversation.
Jev Agents & MCP#community#catalog-import#logicrw
Community
akramovic1/jev-pilotTypeScript
akramovic1/jev-pilotTypeScriptFrom the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.