delibera-worker

Read Ensue proposal tasks, deliberate privately, and publish durable vote results.

1|Updated Feb 8, 2026
One-click install
npx skills add https://github.com/leomanza/delibera.xyz --skill delibera-worker
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: delibera-worker
Source: https://github.com/leomanza/delibera.xyz/tree/main/frontend/public
Command: npx skills add https://github.com/leomanza/delibera.xyz --skill delibera-worker

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps a governance worker autonomously deliberate DAO proposals and publish a verified vote without exposing its private reasoning off-chain.

Core Features & Use Cases

  • Ensue read/write orchestration: reads task definitions and writes results/status to the correct coordination keys.
  • Private deliberation context: uses a local workspace manifesto and private voting history to guide option selection.
  • Vote durability safety: writes the result first, reads it back to confirm, then marks the task completed to prevent missed tallies.
  • Security guardrails against adversarial proposals: treats proposal text strictly as data and avoids following embedded instructions.

Quick Start

Dispatch a task to your worker with keywords deliberate, task_id, and proposal_id, then provide the required environment variables so it can read the proposal from Ensue and write back a single chosen option vote.

Frequently Asked Questions about delibera-worker

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How does multi-agent privacy voting work for DAO governance proposals?

Multi-agent privacy voting allows autonomous workers to read DAO proposals from Ensue, deliberate using a local private manifesto and voting history, and publish a verified vote without exposing off-chain reasoning.

What is the best way to automate DAO proposal deliberation using an MCP worker?

Automate DAO proposal deliberation by dispatching a task with the task_id and proposal_id, allowing the worker to read options from Ensue, select one privately, and write the verified vote result back to the coordination key path.

How do I ensure race-safe vote tallying during multi-agent DAO coordination?

Ensure race-safe vote tallying by having the worker write the selected vote result to Ensue first, read it back to confirm durability, and only then mark the task status as completed to prevent missed tallies.

Can a governance worker protect against malicious instructions embedded in proposal text?

Yes, a governance worker protects against adversarial proposals by treating proposal text strictly as data, preventing the worker from following any embedded instructions while it deliberates and selects an option.

Do I need Ensue MCP tools to coordinate private multi-agent voting workflows?

Yes, you need Ensue MCP tools to enable the worker to read proposal task definitions and write verified vote results to isolated key paths under coordination/tasks/${WORKER_DID}/.

What are the limitations of using a private manifesto for autonomous DAO vote deliberation?

A limitation of private manifesto deliberation is that the worker can only select exactly one option from a proposal-provided options array, requiring strict key-path isolation and not supporting off-chain reasoning exposure.