- Home
- Skills
- Agents & Orchestration
- jroute-cli
Jroute Cli
Local System-1 CLI meta-router for Claude Code & Antigravity (agy) with automatic sub-model/effort selection and git-free .jrc_context.md shared memory.
$7
Works with the AI Agents you already use
jroute-cli
Example session with this skill installed
jroute --json "Refactor the database schema to support multi-tenancy in the Odoo module"
- Read your context and instructions
- Compiled the jroute-cli
The system has analyzed the prompt complexity and production risk. Based on the workspace context, it has selected the optimal agent and reasoning parameters.
Routing Decision Meta-Data
| Parameter | Value |
|---|---|
| Selected Tool | agy (Antigravity) |
| Sub-Model | claude-3-5-sonnet |
| Reasoning Effort | xhigh |
| Complexity Score | 0.92 |
| Risk Level | High |
| Reasoning | Task involves structural database changes and multi-tenancy logic, requiring maximum reasoning depth and architectural awareness. |
{
"routing": {
"agent": "antigravity",
"command": "agy",
"model": "claude-3-5-sonnet",
"effort": "xhigh",
"context_sync": true
},
"analysis": {
"intent": "architectural_refactor",
"entities": ["database", "multi-tenancy", "odoo"],
"safety_check": "passed"
}
}
Next steps
- Run the command without the
--jsonflag to initiate theagysession. - Review
.jrc_context.mdto ensure the Odoo module structure is correctly indexed. - Use
/ctxwithin thejrcREPL to update shared memory after the refactor is complete.
Connects securely to your tools. The creator never sees your data.
About this skill
Manually choosing between CLI agents and reasoning models introduces friction and decision fatigue. You often waste high-tier credits on simple git commits or struggle with low-reasoning models on complex architectural refactors.
What it does
- Intelligent Routing evaluates prompt complexity and production risk to pick between Claude Code and Antigravity.
- Dynamic Calibration selects the optimal sub-model and reasoning effort level (low to xhigh) for every task.
- Shared Memory maintains a git-free
.jrc_context.mdfile to persist task history and architectural notes across different agent sessions. - Context Awareness analyzes the current directory, project files, and tech stack markers to provide the router with workspace metadata.
- Risk Assessment detects high-stakes production tasks to trigger safer tools or higher reasoning models automatically.
How it works
- Initialize the workspace context by running
jrcorjroutein your project root to generate the shared memory layer. - Input your prompt directly via
jroute "task"or enter the interactivejrcREPL for a multi-turn session. - Review the automated routing decision, where the Jev SDK calculates tool fit and model effort based on your local Ollaya server's analysis.
- Execute the task as the router launches the chosen CLI agent with the pre-configured model flags and shared context.
Frameworks & tools
Built with Python and the TypeSafe AI Jev SDK. It integrates directly with Claude Code, Antigravity (agy), and local Ollaya servers for private, local-first decision routing.
Why this beats prompting it yourself
Manual selection is reactive and inconsistent. This skill uses a dedicated System-1 logic layer to make objective decisions on tool fit and reasoning depth before the expensive agent even starts, saving both time and API tokens.
Use cases
- Route simple git operations to low-cost models while sending complex debugging to high-reasoning agents.
- Persist project knowledge between Claude Code and Antigravity sessions using a local memory file.
- Prevent production accidents by automatically escalating reasoning effort for high-risk infrastructure prompts.
- Test routing logic using the
--dry-runor--jsonflags to see model probability distributions.
Known limitations
Requires a local Ollaya server running at the specified port and pre-installed Claude Code or Antigravity binaries. Shared memory relies on MD5 diffing which may be slow in massive directories.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
Requires local Ollaya server at port 11435. Claude Code and Antigravity binaries must be globally accessible. Shared memory MD5 diffing may lag in very large repositories.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 2 days ago
- Software
Python