Claude Skill MartBrowse skillsQuick linesLearn by videoTerminal guideWhat is a Skill?
 
← Back to list

Agentica-Claude Proxy Integration Guide

A developer reference for wiring Agentica agents to a Claude Code CLI proxy, covering required flags, response formats, and common failures.

Dev & CodingAdvanced★ 3,934⑂ 299AI score 7/10Last updated: Jan 26, 2026

What it does

Documents how to build and debug a proxy that lets Agentica SDK agents run through Claude Code CLI (claude -p) with real tool access.

  • Requires --allowedTools Read Write Edit Bash, otherwise Write/Edit fail with "permission denied"
  • SSE streaming response format Agentica expects (data: {...}\n\n, data: [DONE])
  • REPL contract: return results inside a python block with a return statement
  • An anti-hallucination prompt block that forces agents to actually invoke tools before claiming success
  • Sandbox rules: /tmp is blocked, use project-relative paths like workspace/
  • Error → cause → fix table plus health-check and curl test commands

Who it's for

  • Engineers building Agentica swarms that need Claude Code's file and shell tools
  • Anyone writing a local OpenAI-compatible proxy in front of the Claude CLI
  • Users of the Continuous-Claude-v3 repo layout (reference paths assume it)

Examples

  1. Agent reports a file was created but it isn't there → append the anti-hallucination section and verify with ls -la workspace/test.txt.
  2. Agentica throws APIConnectionError → switch the proxy from plain JSON to SSE streaming.
  3. Smoke-test the proxy alone with curl -s http://localhost:8080/v1/chat/completions -d '{...}'.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/parcadei/Continuous-Claude-v3/HEAD/.claude/skills/agentica-claude-proxy/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the .claude/skills/agentica-claude-proxy folder from the GitHub repo parcadei/Continuous-Claude-v3 into my ~/.claude/skills/agentica-claude-proxy/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/parcadei/Continuous-Claude-v3.git /tmp/cc-v3 && mkdir -p ~/.claude/skills && cp -r /tmp/cc-v3/.claude/skills/agentica-claude-proxy ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Clone the repository: git clone https://github.com/parcadei/Continuous-Claude-v3.git
  2. Create the skills directory: mkdir -p ~/.claude/skills
  3. Copy the skill: cp -r Continuous-Claude-v3/.claude/skills/agentica-claude-proxy ~/.claude/skills/
  4. Restart Claude Code and confirm the skill is listed. Note it is marked user-invocable: false, so it is loaded as context rather than called directly.
  5. For real use you also need the repo's scripts/agentica/claude_proxy.py, the uv tool, and the Claude CLI. Start the proxy with uv run python scripts/agentica/claude_proxy.py --port 8080.
  6. Verify with curl http://localhost:8080/health before starting the Agentica server on port 2345.
View source on GitHub ↗License: MIT