interlinked-cli
Local guard layer for coding agents that blocks dangerous tool calls by deterministic rules, with optional Jev judgments.
Overview
interlinked-cli is a local control plane for AI coding agents. It installs hooks into Claude Code and Codex (supported) and experimental adapters for Cursor, Copilot CLI, Gemini CLI, OpenCode and Pi, evaluates each delivered tool call against deterministic policy rules, blocks dangerous ones in milliseconds, and keeps a local, greppable activity log. It needs no remote service; state lives under .interlinked/, and a wizard installs hooks, skills and a daemon, with a documented disable and uninstall path.
It targets teams that need to prove every controllable agent action was checked against policy before it ran. The core block-or-allow decision deliberately has no model in the path so verdicts are explainable. According to the logicrw list, Interlinked adds optional Jev judgments and evidence checks on top of these local checks; the fetched README excerpt does not describe that layer, so its scope is unconfirmed.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Yes / no | Does this agent action or claim meet the policy or evidence requirement?(inferred) | yes · no |
How it works
- Agent emits a tool call through a supported hook
- Deterministic rules allow or block it in milliseconds
- Optionally, Jev judges evidence or softer policy questions
- Every event is logged locally under .interlinked/
Strengths
- Deterministic, explainable blocking with no model in the core path
- Works locally with no remote service
- Supports several coding agents through hooks
Caveats
- Jev integration is optional and not described in the README excerpt
- Non-Claude/Codex adapters are experimental
Getting started
`npm install -g github:QuentinCody/interlinked-cli`, then run `interlinked` inside the repo to start the wizard.At a glance
- Type
- CLI
- Status
- beta
- Language
- TypeScript
- Runtime
- Node CLI, agent hooks, daemon
- Has code
- yes
- License
- MIT
Use cases and integrations
How curated lists describe it
Interlinked adds optional Jev judgments and evidence checks to local coding-agent checks.
logicrw/awesome-jev-projects
Frequently asked questions
Does Interlinked rely on Jev to block commands?
No. Blocking uses deterministic rules; Jev judgments are an optional extra per the list.
Which agents are supported?
Claude Code and Codex are supported; Cursor, Copilot CLI, Gemini CLI, OpenCode and Pi are experimental.
How do I remove it?
`interlinked disable --uninstall` removes a wizard installation.
Related projects
Security hook giving Codex, Copilot, Gemini, Cursor, pi, OpenCode and ACP an auto-mode-like guard via three Jev questions.
LibrarybetaTypeScriptCompiler that turns agent rules in CLAUDE.md and similar files into testable Jev programs: typed questions plus a coded verdict.
ProjectbetaTypeScriptPre-execution hook for coding agents that lets shell commands run, asks a human, or blocks them, with optional Jev classification.
ProjectbetaJavaScriptPython guardrail with a Streamlit dashboard that sends agent tool calls to Jev and allows, blocks or asks for clarification.
AppprototypePythonAgent-ergonomic Jev CLI and PreToolUse hook that scores shell commands for hazards and screens fetched text for injection.
PluginbetaTypeScriptPreToolUse firewall for Claude Code and Codex: shell-aware rules first, then Jev returns allow, ask or block, failing closed.
HookbetaTypeScript