deepu105/pi-automode-classifier
User repository on GitHub
Auto mode for Pi: rules first, then a local decision model, then a confirm prompt
- Forks
- 0
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.
Jev directory
CommunityJev Integrations & Adapters
Pi Automode Classifier
An auto-mode plugin for the Pi coding agent that classifies shell commands before they run: built-in rules decide most calls, and Jev (hosted) or Kev and Laya (local, CPU, offline) route the rest to run, confirm, or block.
- Stars
- 7
- Language
- TypeScript
- License
- MIT
- Reading time
- 2 min
- Published by
- Community
- Author
- deepu105
- Last verified
- Oct 2026
- Languages
- TypeScript
- Platforms
- CLI
- Capabilities
- Open sourceRuns locallyRequires API key
Highlights
- Rules first: rules.ts parses shell commands including $(...), bash -c strings, and sudo, timeout, or uv run wrappers.
- Each tool call gets one of three outcomes: runs, needs a confirm prompt, or is blocked.
- With Kev or Laya on CPU it works fully offline and does not use the GPU; the hosted Jev route is opt-in.
- Built because Pi runs every tool call without approval; the author wanted something like Claude Code's auto mode.
- Not a sandbox: it guards the agent's tool calls, not ! commands you type; extensions run with your permissions.
Watch out
README reviewed 2026-10-17; the digest truncates mid-sentence in the rules section, so the full verdict logic, install steps, and model prompts are not visible. The README itself warns it is not isolation and points to Pi's security docs for real sandboxing.
Questions about this entry
Short answers drawn from the linked project or page.
An auto mode plugin for the Pi coding agent by deepu105 that uses rules plus a decision model — hosted Jev or local Kev and Laya — to classify shell commands before they run, so risky tool calls prompt instead of executing silently.
Jev is one of the classifier options, used hosted through the TypeSafe API. Commands are sent to the provider with Jev, which is why the README calls that route opt-in; Kev and Laya run locally on CPU instead.
Built-in rules parse the shell command — including $(...), bash -c strings, and sudo, timeout or uv run wrappers — and hand back one of four verdicts; the model checks commands the rules do not know, and risky calls get a confirmation prompt.
Yes, with Kev or Laya on CPU: the README says it then works fully offline and does not use the GPU. The hosted Jev route sends commands to the provider, so it is opt-in.
No. The README says it guards the agent's tool calls only: it does not guard ! commands you type, and extensions run with your user's permissions. It points to Pi's security docs for real isolation.
Because Pi runs every tool call without asking for approval, and the author wanted behavior like Claude Code's auto mode: allow the routine calls, prompt on the risky ones, block the dangerous ones.
MIT licensed TypeScript from deepu105, with topics pi-extension and pi-package and 7 stars. The digest truncates before the install section, so check the repository for the exact Pi install command and configuration.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
X: Agent Handoff Gate: agents verify what they hand off
X: Jev as an LLM guardrail inside Agentgateway
X: Vercel's fx safety reviewer, measured at 18× faster
Reddit: Early experiments wiring Jev into Pi as a safety layer and model router.
More Jev resources like this
An opencode plugin (also for pi and Claude Code) that puts every bash command in front of a System One decision model before it runs: a small model scores destructiveness, secret leakage, and your repo's GUARDRAILS.md policy and blocks bad calls.
Jev Integrations & Adapters#community#github#opencode
Community
berget-ai/guardrails-mdTypeScript
berget-ai/guardrails-mdTypeScriptAn OpenCode 2 plugin that sends every shell command — local or over FarHand's remote shell — to Jev first: six typed questions about global installs, host litter, harmful actions and privacy, and a permission prompt whenever something looks risky.
Jev Integrations & Adapters#community#catalog-import#logicrw
Community
cogflux/opencode-jev-guardTypeScript
cogflux/opencode-jev-guardTypeScriptA pi coding-agent extension that compacts long transcripts without rewriting them: Jev answers only keep/kind questions about chunked history, original lines go into a verbatim ledger, and compaction is a deterministic render plus durable topic files. Calls the TypeSafe System One endpoint directly.
Jev Integrations & Adapters#community#catalog-import#logicrw
Permission gate extension for the pi coding agent: deterministic allow/deny rules and protected paths settle clear cases first, then gray-zone tool calls get one structured classify() verdict; Jev is an optional native classifier via TypeSafe direct, OpenRouter, or other gateways, failing closed on error.
Jev Tools & CLIs#community#guardrails#permissions
Community
jesset/pi-verdictTypeScript
jesset/pi-verdictTypeScriptFocus mode for Omarchy and Hyprland that judges each screen against the task you declared and escalates from a red reticle to a window haze and, as a last resort, closing the drifting browser tab. Each check is one Jev on-task classification, with screenshots OCR'd locally and never uploaded.
Jev Integrations & Adapters#community#focus#linux
Community
iYassr/omarchy-laserPython
iYassr/omarchy-laserPythonLocal tool-routing classifier for coding agents: a gateway and MCP tool that ask Jev which tool to call next, keep your agent's own model and login, and log every decision so you can start by observing.
Jev Integrations & Adapters#community#catalog-import#logicrw
Community
felpsdev/jev-classifierTypeScript
felpsdev/jev-classifierTypeScriptFrom 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.