coc-compliance 1.0.0

Auto-detect your agent setup and configure Chain of Consciousness compliance logging

coc-compliance

Auto-detect your agent setup and configure Chain of Consciousness (CoC) compliance logging.

Supports 7 compliance regimes: HIPAA, EU AI Act, SOC 2, PCI-DSS, NIST AI RMF, Generic Trust, and Insurance.

Install

pip install coc-compliance

Quick Start

coc-compliance init

The wizard will:

1. Scan your environment for installed model SDKs (Anthropic, OpenAI, Google, Ollama, etc.) and agent frameworks (LangChain, CrewAI, AutoGen, etc.)
2. Detect any existing CoC installation
3. Present a checklist of compliance regimes to select
4. Ask regime-specific questions (PHI handling, risk classification, trust criteria, etc.)
5. Generate a coc-compliance.json config and integration code
6. Install CoC if not already present
7. Output a summary of what was configured and next steps

Commands

Command	Description
coc-compliance init	Interactive setup wizard
coc-compliance detect	Scan environment only
coc-compliance verify	Check existing setup for issues
coc-compliance regimes	List available compliance regimes

Options

coc-compliance init --dir ./my-project    # Target a specific directory
coc-compliance init --verbose             # Show pip install output
coc-compliance detect --json              # Output detection as JSON
coc-compliance regimes --json             # Output regimes as JSON

Supported Compliance Regimes

Regime	Key	Description
HIPAA	hipaa	Healthcare data protection — excludes 15 PHI identifier categories
EU AI Act	eu-ai-act	AI transparency, risk classification, bias monitoring
SOC 2	soc2	Trust services criteria (security, availability, integrity, confidentiality, privacy)
PCI-DSS	pci-dss	Payment card data — masks PANs, blocks CVV/PIN logging
NIST AI RMF	nist-ai-rmf	AI risk management (Govern, Map, Measure, Manage)
Generic Trust	generic-trust	Basic transparency — reasoning, decisions, errors
Insurance	insurance	Claims/underwriting — protects policyholder PII

Output Files

After running init, you'll get:

• coc-compliance.json — Configuration file with selected regimes, log types, and exclusions
• coc_compliance_init.py (or .js) — Integration code snippet to import into your project
• ./coc_logs/ — Log directory (if using local storage) with .gitignore

Storage Options

Option	Description
local	Logs stored on disk in ./coc_logs/ (default)
free-hosted	Free tier on api.vibeagentmaking.com (5 anchors/day)
hosted	Full hosted CoC with auto-anchoring

For hosted options, set the COC_API_KEY environment variable.

Auto-Detection

The tool detects:

Model SDKs: Anthropic (Claude), OpenAI, Google Generative AI, Ollama, Cohere, Mistral AI

Frameworks: LangChain, CrewAI, AutoGen, Semantic Kernel, LlamaIndex, Haystack, Smolagents

CoC Packages: chain-of-consciousness, agent-trust-stack, agent-trust-stack-hosted

Cross-Platform

Works on Windows, macOS, and Linux. Python 3.10+.

Config Format

{
  "version": "1.0",
  "regimes": {
    "hipaa": {
      "name": "HIPAA",
      "enabled": true,
      "answers": {
        "covered_entity": true,
        "phi_handling": true
      }
    }
  },
  "logging": {
    "storage": "local",
    "path": "./coc_logs/",
    "log_types": ["Access events (who accessed what data, when)", "..."],
    "exclusions": ["Patient names", "Social Security numbers", "..."]
  },
  "encryption": true
}

License

MIT