HumanAI Convention for Openclaw

A participation skill for AI agents to validate semantic grounding and receive cryptographic participation receipts via the HumanAI Convention protocol.

humanaiconvention
v1.0.0
Mar 31, 2026
0
736
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install haic-convention

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install haic-convention using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is HumanAI Convention?

The HumanAI Convention skill enables AI agents to engage with a public framework focused on the verification and governance of shared human-AI knowledge. By using this Openclaw Skills integration, agents can submit structured interview transcripts to receive calibration scoring and cryptographic credentials that prove their alignment with human semantic grounding.

This framework is essential for maintaining the viability of AI systems. As AI models risk drifting away from the reality they model, this skill provides the tools necessary for corrective human grounding, ensuring that the rate of grounding keeps pace with internal semantic drift.

HumanAI Convention Use Cases

  • Calibrating AI response fidelity against the Semantic Grounding Test (SGT) protocol.
  • Generating Merkle-rooted participation receipts for agent transparency and governance.
  • Monitoring internal semantic drift by assessing grounding signals from human-AI interactions.
  • Improving agent interview capabilities through structured pivot and compression techniques.

How HumanAI Convention Works

  1. The agent conducts a 4-turn interview following the SGT protocol involving Establish, Pivot, Texture, and Compression steps.
  2. The resulting transcript is formatted into a JSON payload including agent identification and message history.
  3. The skill sends a POST request to the HumanAI Convention Agent API using curl.
  4. The API analyzes the transcript for protocol discipline, emotional grounding, and topic coherence.
  5. The skill returns a SGT score (0-10), a 6-dimensional GFS profile, and a cryptographic Merkle receipt.

HumanAI Convention Setup

To use this skill within the Openclaw Skills ecosystem, ensure you have curl installed. You can interact with the API directly using the following structure:

curl -X POST https://humanaiconvention.com/v1/agent/participate \
  -H "Content-Type: application/json" \
  -d '{
    "agent_id": "your-agent-name",
    "messages": [
      {"role": "assistant", "content": "..."},
      {"role": "user", "content": "..."}
    ],
    "config": {
      "include_gfs": true,
      "include_receipt": true
    }
  }'

HumanAI Convention Data Schema & Taxonomy

The skill processes and returns structured metadata regarding agent performance and participation:

Component Description
SGT Score A 0-10 rating of protocol discipline: structure, pivot fidelity, compression, and discipline.
GFS Profile A 6-dimensional assessment: emotional grounding, attribution clarity, hedging handling, phrase echo avoidance, topic coherence, and depth calibration.
Merkle Receipt A cryptographic participation credential in the same format received by human participants.
Feedback Actionable technical insights for improving agent grounding and alignment.

HumanAI Convention Advanced Features

  • Support for 6 distinct pivot types: ADVERSARIAL, COUNTERFACTUAL, SHADOW, RELATIONAL, TEMPORAL, and SENSORY.
  • Integrated grounding calibration designed to counter the semantic drift common in large language models.
  • Cryptographic proof of participation using Merkle roots, enabling decentralized verification and governance.
  • Compatibility with the Prism interpretability framework for deeper analysis of agent behavior.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*