Core Concepts¶

Start here to understand the foundational concepts of the Agentic Cognitive Governance Protocol.

Learning Paths¶

New to ACGP?

Start with these essentials:
1. What is ACGP? - Overview and use cases
2. Why ACGP? - Real incidents and ROI
3. ARI Framework - Core governance model
4. Interventions - Response system
Ready to Design?

Design your governance strategy:
1. ACL Tiers - Choose agent autonomy level
2. Reflection Blueprints - Define policies
3. Governance Contracts - Set SLAs
4. Tripwires - Safety limits
Need Advanced Topics?

Dive deeper:
1. CTQ Evaluation - Quality metrics
2. Cognitive Traces - Decision logs
3. Trust System - Dynamic adjustment

Concept Map¶

Understanding how ACGP concepts build on each other:

graph TD
    subgraph foundation [Foundation]
        CT[Cognitive Traces]
        ARI[ARI Framework]
    end

    subgraph governance [Governance Layer]
        ACL[ACL Tiers]
        INT[Interventions]
        CTQ[CTQ Evaluation]
    end

    subgraph advanced [Advanced Features]
        TRIP[Tripwires]
        TRUST[Trust System]
        BP[Blueprints]
        GC[Governance Contracts]
    end

    CT --> ARI
    ARI --> ACL
    ARI --> INT
    ARI --> CTQ
    ACL --> TRUST
    INT --> TRIP
    CTQ --> TRUST
    TRUST --> GC
    BP --> GC
    TRIP --> GC

    click CT "/learn/concepts/cognitive-traces/"
    click ARI "/learn/concepts/ari-framework/"
    click ACL "/learn/concepts/acl-tiers/"
    click INT "/learn/concepts/interventions/"
    click CTQ "/learn/concepts/ctq-evaluation/"
    click TRIP "/learn/concepts/tripwires/"
    click TRUST "/learn/concepts/trust-system/"
    click BP "/learn/concepts/reflection-blueprints/"
    click GC "/learn/concepts/governance-contracts/"

Reading the map:

Start with Cognitive Traces - every decision needs a record
Core framework ARI (Agent-Reflection-Intervention) evaluates traces
Governance ACL tiers, Interventions, and CTQ scoring work together
Advanced features build on the foundation for production systems

Concept Overview¶

Foundation¶

The ACGP governance model is built on three pillars:

1. Agent-Reflection-Intervention (ARI)
Continuous evaluation of agent reasoning and actions with proportionate responses.

2. Access Control Levels (ACL)
Six tiers from read-only (ACL-0) to maximum oversight (ACL-5), matching agent capability and risk.

3. Cognitive Trace Quality (CTQ)
Multi-dimensional scoring of decision quality across reasoning, grounding, ethics, safety, and context.

Key Mechanisms¶

Interventions: Six response levels (OK, Nudge, Escalate, Block, Halt, Flag) provide graduated control.

Tripwires: Proactive safeguards that trigger before quality degradation becomes critical.

Trust System: Dynamic adjustment based on behavior history, enabling agents to earn autonomy.

Governance Contracts: Explicit negotiation of latency/quality trade-offs per action type.

Quick Reference¶

Concept	Purpose	Where Used
ARI Framework	Core governance flow	All conformance levels
ACL Tiers	Agent autonomy classification	Design time
CTQ Evaluation	Decision quality scoring	Runtime evaluation
Interventions	Response actions	Runtime control
Tripwires	Hard safety limits	Runtime protection
Trust System	Behavioral adaptation	Long-term governance
Blueprints	Policy definitions	Configuration
Contracts	Performance SLAs	Standard+ conformance

Next Steps¶

After understanding the concepts:

Implementation Guides Technical Specifications