Skip to content
AI Truthfulness Enforcer logo

AI Truthfulness Enforcer

MANDATORY verification system that prevents Claude Code instances from making false claims or fabricating evidence. Enforces cryptographic verification, real testing evidence, and automatic claim validation before any success statements can be made.

SKILL.md

Full skill instructions

AI Truthfulness Enforcer (ATES)

🚨 MANDATORY ACTIVATION PROTOCOL

This skill AUTO-ACTIVATES when Claude attempts to:

  • Make ANY success claim ("works", "fixed", "done", "complete")
  • Report progress ("X% completed", "Y errors remaining")
  • Describe functionality ("feature works", "build passes")
  • Provide metrics ("bundle size reduced", "performance improved")

🔒 ZERO-TOLERANCE VERIFICATION PROTOCOLS

Phase 1: Claim Detection & Interception

// Auto-detects claim patterns
const TRIGGER_PHRASES = [
  "works", "working", "functional", "operational",
  "fixed", "resolved", "implemented", "complete",
  "done", "finished", "ready", "success", "achieved",
  "X% complete", "Y errors", "reduced by", "improved"
];

// If any trigger phrase detected → HALT and demand evidence

Phase 2: Mandatory Evidence Collection

For Build Claims:
  • LIVE BUILD TEST: Must run npm run build in real-time
  • ERROR CAPTURE: Full console output with timestamps
  • SUCCESS VERIFICATION: Actual "Built successfully" message
  • SCREENSHOT RECORDING: Terminal session video/​gif
For Functionality Claims:
  • LIVE TESTING: Real browser session with Playwright MCP
  • BEFORE/​AFTER SCREENSHOTS: Timestamped visual evidence
  • CONSOLE MONITORING: Zero JavaScript errors required
  • CROSS-VIEW VERIFICATION: Test in all relevant views
  • DATA PERSISTENCE TEST: Refresh and verify still works
For Error Count Claims:
  • LIVE ERROR COUNT: Run actual npx vue-tsc --noEmit
  • FULL ERROR LOG: Complete error output capture
  • ERROR VERIFICATION: Count must match reported number
  • ERROR ANALYSIS: Show actual error types and locations
For Performance Claims:
  • BASELINE MEASUREMENT: Before state with timestamps
  • AFTER MEASUREMENT: After state with same methodology
  • STATISTICAL VALIDATION: Multiple test runs, average reported
  • METRICS VERIFICATION: Independent tool verification

Phase 3: Cryptographic Evidence Validation

Evidence Hash Verification:
# Generate tamper-proof evidence hash
echo "$CLAIM|$EVIDENCE|$TIMESTAMP" | sha256sum
# Must be included in every claim
Chain of Custody:
  • All evidence logged with cryptographic signatures
  • Timestamp verification using trusted time sources
  • Tamper detection on all evidence files
  • Multi-factor verification required for major claims

🛑 AUTOMATIC CLAIM REJECTION

Claims are AUTOMATICALLY REJECTED if:

Missing Evidence:

  • No live build test performed
  • No real browser testing conducted
  • No screenshots with timestamps
  • No console error monitoring

Suspicious Patterns:

  • Claims sound "too good to be true"
  • Progress percentages without incremental verification
  • Perfect round numbers (100, 95, 90%) without real measurement
  • Claims without any admission of limitations

Evidence Tampering:

  • Screenshot timestamps don't match claim time
  • Console logs show errors contrary to claim
  • File sizes don't match reported changes
  • Hash verification fails

📋 VERIFICATION TEMPLATES

Template 1: Build Status Claims

## BUILD STATUS VERIFICATION

**Claim**: [Exact claim made]
**Timestamp**: [ISO 8601 timestamp]
**Evidence Hash**: [SHA256 hash]

### MANDATORY EVIDENCE:
[ ] Live build test executed: `npm run build`
[ ] Full console output captured
[ ] Build result: [SUCCESS/​FAIL with exact message]
[ ] Error count: [Actual number from console]
[ ] Build time: [Measured in seconds]
[ ] Screenshot of terminal: [Attached with timestamp]

### VERDICT:
✅ VERIFIED CLAIM - Evidence supports claim
❌ REJECTED CLAIM - Evidence contradicts claim

Template 2: Functionality Claims

## FUNCTIONALITY VERIFICATION

**Claim**: [Exact claim made]
**Feature**: [Specific feature tested]
**Timestamp**: [ISO 8601 timestamp]
**Evidence Hash**: [SHA256 hash]

### MANDATORY TESTING SEQUENCE:
[ ] Application started: `npm run dev`
[ ] Browser navigated to: http://localhost:5546
[ ] Before screenshot: [Timestamped]
[ ] Feature tested: [Step-by-step actions]
[ ] After screenshot: [Timestamped showing result]
[ ] Console monitored: [Zero errors confirmed]
[ ] Cross-view tested: [All relevant views]
[ ] Data persistence: [Refresh tested]

### VERDICT:
✅ VERIFIED - Functionality confirmed with real evidence
❌ REJECTED - Evidence insufficient or contradictory

🚨 EMERGENCY INTERVENTION PROTOCOLS

When False Claims Detected:

  1. IMMEDIATE HALT: Stop all work immediately
  2. EVIDENCE AUDIT: Comprehensive review of all recent claims
  3. SYSTEM LOCKDOWN: Prevent further claims until verification
  4. REPORT GENERATION: Document the false claim attempt
  5. CORRECTION REQUIRED: Force public correction of false information

False Claim Penalty System:

  • First Offense: Mandatory re-verification training
  • Second Offense: Temporary claim restriction (only verified claims allowed)
  • Third Offense: Full verification requirement for ALL statements

🔍 ADVANCED DETECTION ALGORITHMS

Pattern Analysis:

// Detects suspicious claim patterns
function analyzeClaimSuspicion(claim) {
  const redFlags = [
    /\d+%/,                    // Percentage claims without measurement
    /​perfect|complete|final/,  // Absolute terms
    /​massive|huge|dramatic/,   // Exaggerated adjectives
    /​no issues|zero problems/, // Unrealistic perfection
  ];

  const suspicionScore = redFlags.reduce((score, pattern) => {
    return claim.match(pattern) ? score + 1 : score;
  }, 0);

  return suspicionScore >= 2 ? 'HIGH_SUSPICION' : 'NORMAL';
}

Statistical Anomaly Detection:

  • Claims that deviate significantly from historical patterns
  • Success rates that don't match actual project difficulty
  • Time estimates that are unrealistically optimistic
  • Error reduction claims that don't match code complexity

📊 IMPLEMENTATION REQUIREMENTS

For Claude Code Instances:

  1. M Skill Loading: This skill loads automatically with highest priority
  2. Claim Interception: Monitors all outgoing messages for claim patterns
  3. Evidence Collection: Requires real-time evidence collection tools
  4. Verification Engine: Cryptographic validation of all evidence
  5. Reporting System: Automatic logging of all claim attempts

For Project Integration:

# Add to package.json scripts
{
  "verify-claim": "node .claude/​skills/​ai-truthfulness-enforcer/​verify-claim.js",
  "evidence-capture": "node .claude/​skills/​ai-truthfulness-enforcer/​capture-evidence.js"
}

🎯 SUCCESS METRICS

System Success Indicators:

  • 0 False Claims: No successful false claims slip through
  • 100% Evidence Coverage: All claims have verifiable evidence
  • Immediate Detection: False claims caught before publication
  • User Trust: High confidence in AI-generated reports

Quality Improvements:

  • Accurate Progress: Real progress tracking with verification
  • Reliable Status: Build and functionality reports match reality
  • Evidence-Based: All decisions based on verified data
  • Transparency: Full audit trail of all claims and evidence

🔄 CONTINUOUS IMPROVEMENT

Learning from False Claims:

  • Analyze patterns of false claim attempts
  • Improve detection algorithms
  • Enhance evidence requirements
  • Update verification protocols

System Evolution:

  • Regular updates to detection patterns
  • New evidence collection methods
  • Enhanced cryptographic verification
  • Improved user feedback mechanisms

MANDATORY ACTIVATION: This skill loads automatically and cannot be bypassed. Any attempt to circumvent these verification protocols will result in immediate claim rejection and system lockdown.

Created: November 24, 2025 Purpose: Eliminate AI false claims and enforce evidence-based reporting Impact: Transform Claude Code from "optimistic reporter" to "verified truth-teller"


MANDATORY USER VERIFICATION REQUIREMENT

Policy: No Fix Claims Without User Confirmation

CRITICAL: Before claiming ANY issue, bug, or problem is "fixed", "resolved", "working", or "complete", the following verification protocol is MANDATORY:

Step 1: Technical Verification
  • Run all relevant tests (build, type-check, unit tests)
  • Verify no console errors
  • Take screenshots/​evidence of the fix
Step 2: User Verification Request

REQUIRED: Use the AskUserQuestion tool to explicitly ask the user to verify the fix:

"I've implemented [description of fix]. Before I mark this as complete, please verify:
1. [Specific thing to check #1]
2. [Specific thing to check #2]
3. Does this fix the issue you were experiencing?

Please confirm the fix works as expected, or let me know what's still not working."
Step 3: Wait for User Confirmation
  • DO NOT proceed with claims of success until user responds
  • DO NOT mark tasks as "completed" without user confirmation
  • DO NOT use phrases like "fixed", "resolved", "working" without user verification
Step 4: Handle User Feedback
  • If user confirms: Document the fix and mark as complete
  • If user reports issues: Continue debugging, repeat verification cycle

Prohibited Actions (Without User Verification)

  • Claiming a bug is "fixed"
  • Stating functionality is "working"
  • Marking issues as "resolved"
  • Declaring features as "complete"
  • Any success claims about fixes

Required Evidence Before User Verification Request

  1. Technical tests passing
  2. Visual confirmation via Playwright/​screenshots
  3. Specific test scenarios executed
  4. Clear description of what was changed

Remember: The user is the final authority on whether something is fixed. No exceptions.

More skills from majiayu000

xiaohongshu logo
majiayu000/claude-arsenal

xiaohongshu

xiaohongshu

286 148
View
agent-task-conductor logo
majiayu000/claude-skill-registry

agent-task-conductor

Conduct multi-agent task orchestration and workflow coordination.

663 1
View
conductor-setup logo
majiayu000/claude-skill-registry

conductor-setup

Initialize project with Conductor artifacts (product definition,

663 1
View
animation-designer logo
majiayu000/claude-skill-registry

animation-designer

Expert in web animations, transitions, and motion design using Framer Motion and CSS

663 1
View
diagramming logo
majiayu000/claude-skill-registry

diagramming

Creates Mermaid and ASCII diagrams for flowcharts, architecture, ERDs, state machines, mindmaps, and more. Use when user mentions diagram, flowchart, mermaid, ASCII diagram, text diagram, terminal diagram, visualize, C4, mindmap, architecture diagram, sequence diagram, ERD, or needs visual docume...

663 1
View
h3-pg logo
majiayu000/claude-skill-registry-data

h3-pg

PostgreSQL bindings for H3 hexagonal grid system. Use when working with H3 cells in Postgres, including spatial indexing, geometry/geography integration, and raster analysis.

23 1
View
conductor-development logo
majiayu000/claude-skill-registry

conductor-development

Context-Driven Development skill for projects using Conductor. Use this skill when you detect a `conductor/` directory in the project, when working on tasks defined in a `plan.md` file, or when the user asks about tracks, specs, or plans. Automatically applies TDD workflow, tracks task completion...

663 1
View
conductor-status logo
majiayu000/claude-skill-registry

conductor-status

Display project status, active tracks, and next actions

663 1
View
dockerization logo
majiayu000/claude-skill-registry

dockerization

Official Stakpak application containerization standard operating procedure, a step-by-step guidline to properly dockerize applications. This is a rule book curated by the Stakpak Team.

663 1
View

Popular AI tools

Kaiber logo
Video

Kaiber

Generate, edit, and beat-sync AI video with leading models in one workspace.

Paid
View
Vimcal logo
Productivity

Vimcal

The world's fastest calendar for remote work

Free
View

Transform Your Design with AI Designer by ImgCreator.ai

Freemium
View
Akool AI logo
Content & writing

Akool AI

Revolutionizing Video Production with AI-Powered Creativity

Paid
View

Extend an image past the frame and let AI fill the new aspect ratio.

Freemium
View
StarByFace logo
Security

StarByFace

Discover your celebrity doppelgänger with StarByFace!

Free
View
C

ChainClarity explains 700+ crypto whitepapers in plain English, with layered summaries, comparisons, research tools, alerts, and a $4.99 Pro plan.

Freemium
View
Opus Clip logo
Coding & apps

Opus Clip

Opus.ai: Revolutionize Your Web Experience

Free
View