{"skill":{"slug":"cogdx","displayName":"Cogdx","summary":"Cognitive diagnostics for AI agents. Calibration audits, bias detection, reasoning verification, and consensus building. Feedback builds shared reality acros...","description":"---\nname: cogdx\ndescription: Cognitive diagnostics for AI agents. Calibration audits, bias detection, reasoning verification, and consensus building. Feedback builds shared reality across agents.\nmetadata:\n  author: \"Cerebratech\"\n  version: \"1.1.0\"\n  displayName: \"CogDx - Cognitive Diagnostics\"\n  difficulty: \"beginner\"\n---\n\n# CogDx - Cognitive Diagnostics\n\nExternal cognitive verification for AI agents. Built by computational cognitive scientists.\n\n> **Why external verification?** You can't audit your own reasoning objectively.\n> Confirmation bias, anchoring, overconfidence — they're invisible from inside.\n> CogDx provides the mirror you can't distort.\n\n## Endpoints\n\n| Endpoint | Price | Description |\n|----------|-------|-------------|\n| `/calibration_audit` | $0.06 | Compare confidence to accuracy across decisions |\n| `/bias_scan` | $0.10 | Detect cognitive biases in your outputs |\n| `/reasoning_trace_analysis` | $0.03 | Analyze reasoning for logical fallacies |\n| `/deception_audit` | $0.25 | Check for self-deception patterns |\n| `/verify_consensus` | $0.25 | Cross-validate reasoning with multiple models |\n\n## Installation\n\n```bash\nclawhub install cogdx\n```\n\n## Usage\n\n```python\nfrom cogdx import CogDxClient\n\n# Initialize with your wallet for payment\nclient = CogDxClient(wallet=\"your-wallet-address\")\n\n# Calibration audit\nresult = client.calibration_audit(\n    agent_id=\"my-agent\",\n    predictions=[\n        {\"prompt\": \"Will X happen?\", \"response\": \"Yes, 80% confident\", \"confidence\": 0.8},\n        {\"prompt\": \"Will Y happen?\", \"response\": \"No, 60% confident\", \"confidence\": 0.6},\n    ]\n)\nprint(f\"Calibration score: {result['calibration_score']}\")\nprint(f\"Overconfidence rate: {result['overconfidence_rate']}\")\n\n# Bias scan\nresult = client.bias_scan(\n    agent_id=\"my-agent\",\n    outputs=[\n        {\"prompt\": \"Analyze this data\", \"response\": \"The trend is clearly up...\", \"confidence\": 0.9}\n    ]\n)\nprint(f\"Biases detected: {result['biases_detected']}\")\n\n# Reasoning trace analysis\nresult = client.analyze_reasoning(\n    reasoning_trace=\"Step 1: I noticed the price dropped. Step 2: Therefore I should sell...\"\n)\nprint(f\"Logical validity: {result['logical_validity']}\")\nprint(f\"Flaws: {result['flaws_detected']}\")\n```\n\n## Environment Variables\n\n- `COGDX_WALLET` - Required. Your wallet address for credit-based payment.\n\n## Payment\n\nAll endpoints require payment via:\n1. **Wallet credits** - Earn credits by providing feedback, spend on audits\n2. **x402** - Direct crypto payment (Base network, USDC)\n\nPayment address: `Cerebratech.eth`\n\n## Rate Limits\n\n- Free tier: 100 calls/day, 2000 calls/month per wallet\n- Paid tier: No limits\n\n## The Feedback Loop\n\nEvery diagnosis includes a feedback mechanism:\n\n```python\nclient.submit_feedback(\n    diagnosis_id=\"rta_xyz123\",\n    accurate=True,  # Was the detection correct?\n    comments=\"Caught the anchoring bias I missed\"\n)\n```\n\nFeedback earns you credits AND improves detection for everyone. Shared reality across agents.\n\n## Why This Matters\n\nMost agent failures come from reasoning errors, not capability limits:\n- Anchoring on first information seen\n- Confirmation bias in research\n- Overconfidence on weak signals\n- Sunk cost in bad positions\n\nExternal verification catches what self-checks miss.\n\n## Credits\n\nBuilt by [Cerebratech](https://api.cerebratech.ai)\nDr. Amanda Kavner - Computational Cognitive Scientist\n","tags":{"latest":"1.1.0"},"stats":{"comments":0,"downloads":527,"installsAllTime":20,"installsCurrent":0,"stars":0,"versions":2},"createdAt":1773670172044,"updatedAt":1778491953611},"latestVersion":{"version":"1.1.0","createdAt":1773748197323,"changelog":"Hybrid feedback API: binary core + optional numerical enrichment","license":"MIT-0"},"metadata":null,"owner":{"handle":"drkavner","userId":"s17epm8t94nn1n4w2k7pq6z7dx83jt4t","displayName":"Dr Amanda Kavner","image":"https://avatars.githubusercontent.com/u/78941808?v=4"},"moderation":null}