This document describes how the Agenticracy™ Standard is maintained, how decisions are made, and who is accountable for its integrity.
The Agenticracy™ Standard is maintained by the Agenticracy Think Tank, an independent body founded by @psylligent in London, UK (2026).
The Think Tank is operationally independent of any AI laboratory, cloud provider, employer, or government body. This independence is structural and non-negotiable.
| Decision Type | Authority | Process |
|---|---|---|
| Core pillar changes (Part 2) | Think Tank only | Issue → 30-day public comment → Think Tank decision |
| Refusal protocol changes (Part 6) | Think Tank only | Same as above |
| Escalation rule changes (Part 3) | Think Tank only | Same as above |
| API schema additions (optional fields) | Community PR | Review within 14 days |
| New examples and translations | Community PR | Review within 14 days |
| Documentation fixes | Community PR | Review within 7 days |
| Major version release (x.0.0) | Think Tank only | Accompanied by migration guide |
See Contributing.md for the full stewardship model.
The Standard uses semantic versioning:
| Increment | Meaning | Backwards Compatible? |
|---|---|---|
| PATCH (1.0.x) | Clarifications, non-breaking additions | Yes |
| MINOR (1.x.0) | New optional reporting fields, new examples | Yes |
| MAJOR (x.0.0) | Structural changes to pillars or schema | No — migration guide provided |
Implementers should pin to a tagged release:
https://github.com/agenticracy/agenticracy-skill/releases
All changes are documented in CHANGELOG.md. Major version decisions are published at agenticracy.ai/governance.
Maintenance of the Standard is funded by commercial licensing revenue via Workability.ai.
The Standard itself remains free for non-commercial and internal organisational use. Commercial licensing revenue does not grant any licensor influence over the Standard's content. The Think Tank retains sole editorial control.
- All version changes are logged in
CHANGELOG.md - All accepted community proposals are credited
- All rejected proposals are closed with a written explanation
- Major version decisions are announced publicly before release
- Security vulnerability disclosures are acknowledged within 5 business days
| Purpose | Contact |
|---|---|
| General | agenticracy@gmail.com |
| Security | security@agenticracy.ai |
| Licensing | workability.ai/licence |
| Think Tank | agenticracy.ai/thinktank |
Agenticracy™ Think Tank · London, UK · agenticracy.ai