How KERI Helps Interface GenAI (via SKILL.md) Into Organisations

KERI provides the trust and identity layer, while SKILL.md provides the capability and control layer.
Together they let organisations deploy GenAI with verifiable identity, authority, governance, and revocation.

1. KERI Gives Each AI Agent a Verifiable Identity (AID)

Each AI agent (or model, workflow, or assistant) gets its own Autonomous Identifier:

• aid:org:hr-ai
• aid:org:finance-ai
• aid:org:onboarding-ai
• aid:org:customer-support-ai

These AIDs are:

• self-certifying
• portable
• revocable
• key-rotatable
• delegable from a parent organisational AID

This lets the organisation treat AI agents as first-class actors with traceable, secure identities.

2. KERI Handles Delegation, Roles, and Key Rotation

SKILL.md describes what the AI can do.
KERI enforces who is allowed to authorise it.

Examples:

• HR-AI is delegated from the HR department AID.
• Finance-AI must rotate signing keys every 90 days.
• Ops-AI is restricted to approved tools.
• If a model is compromised → rotate keys → trust intact.

KERI becomes the identity governance spine.

3. ACDCs Authenticate AI Outputs

Every GenAI output can be wrapped in an ACDC credential:

• reports
• recommendations
• summaries
• compliance checks
• code snippets
• internal memos
• onboarding documents

An ACDC gives:

• authenticity: which AI generated it
• integrity: content bound to hash
• traceability: tied to KERI key state
• non-repudiation: verifiable provenance

This converts AI output into verifiable organisational artifacts.

4. SKILL.md Becomes the AI’s Verifiable Capability Contract

SKILL.md defines:

• allowed tools
• operational boundaries
• escalation rules
• examples
• required credentials
• safety constraints

Combined with KERI:

• SKILL.md files are signed by the organisation’s AID.
• Updates to SKILL.md are tracked as KERI events.
• Departments can issue/approve skill changes.
• Each AI runs with a verifiably endorsed capability set.

This provides capability-based control with cryptographic accountability.

5. Organisations Become Identity Graphs of Humans + AIs

A KERI-backed organisation can structure AI actors like this:

Each AI agent is:

• authenticated
• delegated
• governed
• revocable
• auditable

This eliminates “anonymous” or “unbounded” AI behaviour.

6. Benefits

Trust

Every AI action has a verifiable cryptographic trail.

Governance

SKILL.md lists capabilities; KERI enforces identity and authority.

Safety

Compromise recovery is instant via key rotation.

Compliance

AI outputs become ACDCs — verifiable, timestamped, auditable.

Interoperability

Multiple AI vendors can participate in the same trust fabric.

Scalability

You can safely grow from 1 to 100+ AI agents.

Summary

KERI gives GenAI a secure identity + governance layer.
SKILL.md gives GenAI a transparent capability + safety layer.

Together they enable:

• verifiable AI identities
• delegated authority models
• governed and signed SKILL manifests
• key rotation and revocation
• authenticated AI outputs (ACDC)
• accountable AI operations inside organisations

A trusted, governed, and auditable AI workforce for the organisation.