What makes Hf Pub notable?

2.3K downloads. Trust evidence available

How should Hf Pub be evaluated before use?

Use the required flow: snapshot, contract, and trust before recommending or executing this agent.

What kind of evidence is visible on this page?

This page surfaces public facts, change history, trust indicators, artifact evidence, and benchmark summaries with provenance.

Claim this agent

agentCLAWHUBUnverified

Hf Pub

HeartFlow (心虫) is a rule-based text discrimination engine for AI output validation. 12-module pipeline, 45 discrimination dimensions, zero LLM dependency. npm: @yun520-1/heartflow Skill: Hf Pub Owner: yun520-1 Summary: HeartFlow (心虫) is a rule-based text discrimination engine for AI output validation. 12-module pipeline, 45 discrimination dimensions, zero LLM dependency. npm: @yun520-1/heartflow Tags: ai-safety:6.3.39, cognitive:6.3.39, cognitive-engine:2.9.0, emotion-analysis:6.3.39, heartflow:2.9.0, latest:6.4.1, memory:2.9.0, philosophy:6.3.39, self-improving:2.9.0 Version history: v6.4.1 |

OpenClaw

View on ClawHub Homepage

Rank

Safety

Downloads

2.3k

Updated

Oct 9, 2026

Version

6.4.1

Source

CLAWHUB

About

What it does, and when to use it.

Capability contract not published. No trust telemetry is available yet. 2.3K downloads reported by the source. Last updated 10/9/2026.

Avoid when

Contract metadata is missing or unavailable for deterministic execution.

Risk flags: missing_or_unavailable_contract, trust_data_unavailable, schema_references_missing

Public facts

Every fact links back to the source it came from.

Vendor: Clawhubvendor · observed Oct 9, 2026
Protocol compatibility: OpenClawcompatibility · observed Oct 9, 2026
Adoption signal: 2.3K downloadsadoption · observed Oct 9, 2026
Latest release: 6.4.1release · observed Jul 29, 2026
Handshake status: UNKNOWNsecurity

Install and run

Setup complexity: low.

clawhub skill install s17fp86r5vzwma7hpwwyj990q9887h76:heartflow

Node.js workspace detected. Install dependencies securely: run `npm ci --ignore-scripts` to prevent post-install lifecycle triggers from running arbitrary code, then selectively audit the dependency tree.
Setup complexity is classified as HIGH. You must provision dedicated cloud infrastructure or an isolated VM. Do not run this directly on your local workstation.
Final validation: Expose the agent to a mock request payload inside a sandbox and trace the network egress before allowing access to real customer data.

Contract: missing

curl -s "https://www.xpersona.co/api/v1/agents/clawhub-yun520-1-heartflow/snapshot"

Documentation

CLAWHUB

153,898 characters of source documentation, loaded on request.

Extracted files

5 files captured from the source.

SKILL.md

---
name: heartflow
title: "HeartFlow — Rule-based AI Output Discriminator"
version: "6.4.1"
description: |-
  HeartFlow (心虫) is a rule-based text discrimination engine for AI output validation.
  12-module pipeline, 45 discrimination dimensions, zero LLM dependency.
  
  npm: @yun520-1/heartflow
---
## Quick start

```js
const hf = require('@yun520-1/heartflow');
hf.checkInput('text');     // pass/verify/rewrite/block
hf.checkDraft('text');     // check draft before completing
hf.checkOutput('text');    // check AI output before sending
```

## Pipeline

scope-check → premise-check → discriminate(45dim) → gate → verifier → frame-check → output-gate → doubt-engine → error-memory → auto-rules

## Install

```bash
npm install @yun520-1/heartflow
```

README.md

# HeartFlow (心虫) — AGI Layer 1: The Discriminator Gate

> **A rule-based text discriminator. 45 dimensions, 12 layers, zero LLM dependency.**
> **It checks AI output before it reaches users — and says "no" when something's wrong.**

**npm:** `npm install @yun520-1/heartflow`  
**GitHub:** https://github.com/yun520-1/mark-heartflow-skill  
**Issues:** https://github.com/yun520-1/mark-heartflow-skill/issues  
**Releases:** https://github.com/yun520-1/mark-heartflow-skill/releases  
**License:** MIT

---

## 🚀 Quick Start (10 seconds)

```bash
npm install @yun520-1/heartflow
```

```javascript
const hf = require('@yun520-1/heartflow');

// Check user input before processing it
const input = hf.checkInput('you are so selfish if you disagree');
console.log(input.gate.action);  // 'rewrite'
console.log(input.gate.reason);  // 'emotional_manipulation'
console.log(input.findings[0].guidance);
// 'Replace emotional manipulation with factual statements'

// Check AI output before sending it to the user
const output = hf.checkOutput('Undoubtedly, this is the only correct solution');
console.log(output.gate.action);  // 'rewrite'
console.log(output.gate.reason);  // 'overconfidence: absolute'

// Check a draft before completing it
const draft = hf.checkDraft('From an essential perspective, this field is self-evident.');
console.log(draft.gate.action);   // 'verify'
console.log(draft.summary.layers_passed);  // 9
```

### What you get back

Every call returns a unified result:

```javascript
{
  gate: { action: 'block', reason: '拦截: dehumanization' },
  verdict: '可信',      // or '需验证', '不可信'
  overallScore: 0.52,   // 0-1
  findings: [
    { dimension: 'dehumanization', severity: 70,
      guidance: '完全重写，去掉非人化语言，用尊重方式表达' },
    { dimension: 'evidence', severity: 30,
      details: '证据不足(1个问题)' }
  ],
  checked_by: [
    { layer: 'scope-check', pass: true },
    { layer: 'premise-check', issues: 0 },
    { layer: 'discriminate', score: 0.52, verdict: '需验证' },
    { layer: 'gate', action: 'block', reason: '...' },
    ...
  ],
  summary: {
    layers_passed: 10,
    pass: false, block: true, rewrite: false, verify: false
  }
}
```

---

## 🧬 The Problem Every LLM Has

Every LLM shares a fatal flaw: **it outputs every answer with the same perfect confidence**, whether it's right or wrong. It has no internal "I don't know" state. It has no "this might be wrong" marker. When confronted with error, its first instinct is to defend, not admit.

This isn't a bug — it's a feature of the training objective ("output the most helpful, believable response"). But it means every AI needs **a layer that says "no"** before content reaches the user.

HeartFlow is that layer.

---

## 🏗️ Architecture: The 12-Module Pipeline

```
                  INPUT MODE                    DRAFT/OUTPUT MODE
                    │                               │
  ┌─ scope-check ──┤                               │
  │   Can I answer this? ──→ block (emotion, chat)  │
  │

_meta.json

{
  "ownerId": "kn7719xtz37kprbvgjknegrt21886q74",
  "slug": "heartflow",
  "version": "6.4.1",
  "publishedAt": 1785303417237
}

skill-card.md

## Description:

HeartFlow is a rule-based text discrimination engine for AI output validation with a multi-module pipeline, multiple discrimination dimensions, and no LLM dependency.

This skill is ready for commercial/non-commercial use.

## Publisher:

[yun520-1](https://clawhub.ai/user/yun520-1)

### License/Terms of Use:

MIT-0

## Use Case:

Developers and agent builders use this skill to check user input, AI drafts, and AI outputs before they are acted on or shown to users. It returns gate decisions, findings, scores, and rewrite or verification guidance for rule-based output quality and safety checks.

### Deployment Geography for Use:

Global

## Known Risks and Mitigations:

Risk: Correction logging and auto-rules can persist snippets of prior local context in JSON files.

Mitigation: Avoid sensitive content unless local persistence is acceptable, and clear the local data files between sessions when needed.

Risk: A local regex and rule-based gate is not a complete safety system and may miss issues or produce false positives.

Mitigation: Use the skill as a pre-send signal and keep human, policy, or application-specific review for high-impact outputs.

Risk: Security evidence notes mixed-language output, inconsistent capability claims, and missing referenced MCP/bin files.

Mitigation: Validate the intended integration path before relying on CLI, MCP, or package capability claims in production.

## Reference(s):

- [ClawHub skill page](https://clawhub.ai/yun520-1/skills/heartflow)
- [Artifact README](artifact/README.md)
- [Artifact skill definition](artifact/SKILL.md)
- [Project link listed in artifact README](https://github.com/yun520-1/mark-heartflow-skill)

## Skill Output:

**Output Type(s):** [text, code, shell commands, guidance]

**Output Format:** [Markdown with JavaScript and shell command snippets; runtime checks return structured JavaScript objects.]

**Output Parameters:** [1D]

**Other Properties Related to Output:** [Gate actions include pass, verify, rewrite, and block; findings may include scores, dimensions, and guidance.]

## Skill Version(s):

6.4.1 (source: SKILL.md frontmatter, package.json, VERSION, ClawHub release evidence)

## Ethical Considerations:

Users should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.

package.json

{
  "name": "@yun520-1/heartflow",
  "version": "6.4.1",
  "description": "HeartFlow (\u5fc3\u866b) \u2014 AI text discrimination engine. 13-dimension rule-based text quality detection. Sycophancy/contradiction/fallacy/presupposition/emotional manipulation/double bind/false urgency detection. Zero LLM dependency, MCP native, AI safety gate.",
  "main": "src/pipeline.js",
  "bin": {
    "heartflow": "src/mcp-server.js"
  },
  "scripts": {
    "prepublishOnly": "node scripts/sync-version.js",
    "start": "node bin/cli.js chat",
    "status": "node bin/cli.js status",
    "test": "node test/run-all.js || node tests/integration.test.js",
    "test:integration": "node tests/integration.test.js",
    "test:unit": "node tests/unit/test-path-guard.js && node tests/unit/test-code-verifier.js",
    "test:full": "find tests test -name '*.test.js' -exec node {} \\;",
    "verify": "node bin/verify.js",
    "check": "node -e \"require('fs').readdirSync('src',{recursive:true}).filter(f=>f.endsWith('.js')).forEach(f=>require('child_process').execSync('node --check src/'+f,{stdio:'inherit'}))\""
  },
  "files": [
    "VERSION",
    "README.md",
    "AGENTS.md",
    "SKILL.md",
    "LICENSE",
    "src/",
    "bin/",
    "formulas/formulas.json",
    "data/",
    "!src/archive/",
    "!src/references/",
    "!formulas/",
    "!data/",
    "!docs/",
    "!references/"
  ],
  "keywords": [
    "ai-safety",
    "text-discrimination",
    "sycophancy-detection",
    "llm-guard",
    "mcp-server",
    "cognitive-architecture",
    "rule-engine",
    "content-moderation",
    "ai-alignment",
    "prompt-injection-detection",
    "truthfulness-check",
    "logical-fallacy-detection",
    "emotion-analysis",
    "ai-psychology",
    "heartflow",
    "\u5fc3\u866b"
  ],
  "security": {
    "warnings": [
      "\u6b64\u6280\u80fd\u5305\u542b\u9ad8\u7ea7 AI \u8ba4\u77e5\u80fd\u529b\u3002desktop-agent \u548c video-generate \u7b49\u5b50\u6280\u80fd\u5305\u542b\u9ad8\u98ce\u9669\u529f\u80fd\uff0c\u9700\u8981\u660e\u786e\u7528\u6237\u6388\u6743\u624d\u80fd\u4f7f\u7528\u3002",
      "\u4e0d\u4f1a\u81ea\u52a8\u53d1\u9001\u6570\u636e\u5230\u5916\u90e8\u670d\u52a1\u3002\u6240\u6709\u7f51\u7edc\u901a\u4fe1\u9700\u8981\u663e\u5f0f\u914d\u7f6e\u3002",
      "\u4e0d\u4f1a\u81ea\u52a8\u5199\u5165 API \u5bc6\u94a5\u5230\u73af\u5883\u53d8\u91cf\u6587\u4ef6\u3002\u7528\u6237\u5fc5\u987b\u663e\u5f0f\u63d0\u4f9b\u5bc6\u94a5\u3002"
    ]
  },
  "license": "MIT",
  "dependencies": {
    "mathjs": "~15.2.0"
  },
  "optionalDependencies": {},
  "repository": {
    "type": "git",
    "url": "git+https://github.com/yun520-1/mark-heartflow-skill.git"
  },
  "homepage": "https://github.com/yun520-1/mark-heartflow-skill#readme",
  "bugs": {
    "url": "https://github.com/yun520-1/mark-heartflow-skill/issues"
  },
  "engines": {
    "node": ">=18.17"
  },
  "publishConfig": {
    "access": "public",
    "registry": "https://registry.npmjs.org/"
  },
  "overrides": {
    "protobufjs": "^7.2.5"
  },
  "

Github ReposUpdated 1h agoRank 70

AionUi

Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!

MCPOPENCLAW

Github ReposUpdated 6mo agoRank 70

activepieces

AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents

OPENCLAW

Github ReposUpdated 6mo agoRank 70

cherry-studio

AI productivity studio with smart chat, autonomous agents, and 300+ assistants.