Coaching Trust Framework
Version 1.0 — April 2026
The common good prevails over any organization, including any organization that might one day emerge to steward this standard.
Author: Maxime Bui License: Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0) Status: Public release, version 1.0 Canonical URL: https://coaching-trust-framework.org/document/ctf-v1.0 Repository: https://github.com/macsimplex/coaching-trust-framework Citation:
Maxime Bui (2026). Coaching Trust Framework (CTF), version 1.0. Licensed under CC BY-SA 4.0.
Note on this document
This file is the front matter and structural skeleton of the generic CTF normative document v1.0. The full text is held in the founder’s working files and will be merged into this file before public release.
The structure below reflects the deliberately minimal posture of the generic CTF core. It contains only what is structurally true of any AI coaching agent regardless of professional tradition. Substantive ethical principles specific to coaching traditions live in federation adaptations under ../declinations/ and may be promoted into this generic core over time, by convergence and by formal upstream proposals (see ../GOVERNANCE.md, section 5).
If you are looking for the substantive ethical content of CTF as it applies to a specific coaching tradition, look at the adaptations:
Table of contents
Preamble
Part I — Foundations
- What CTF generic is
- What CTF generic is not
- The inductive posture
- Relation to federation adaptations
Part II — Operational principles (the minimal core)
- Sollicited tool, not conversational companion
- Non-anthropomorphization
- Non-substitution to human professionals
- Auditability by construction
- Anonymization by construction of audit material
- Hierarchy of principles in case of conflict
Part III — The quality cycle (operational requirements for platforms)
- Use of CTF Designer at design phase
- Use of CTF Test Plan and CTF Robustness before deployment
- Use of CTF Audit during deployment, on documented periodicity
- Tripartite supervision (eggspert + certified supervisor + CTF Audit)
- Material maintenance (design documents, test plans, audit reports)
Part IV — Conformance and certification
- Self-declaration: minimal commitments
- Adaptation pathway: relationship to federation adaptations
- Loss of conformance
Part V — Vocabulary
- Defined terms (refer to glossary)
- Rejected terms
Part VI — Revision rules
- Versioning
- Founding principle as upper boundary
- Upstreaming from adaptations into the generic core
Preamble
[To be integrated from the founder’s working files. The preamble states the political position of CTF, its inductive posture, its relationship to federation adaptations, and the spirit in which it is offered to the community.]
Part I — Foundations
1. What CTF generic is
The generic CTF is the structural common floor that every AI coaching agent should satisfy, regardless of the professional tradition it serves. It defines the minimal operational grammar — posture, the quality cycle, auditability, anonymization — and nothing else.
2. What CTF generic is not
The generic CTF is not the full ethical framework for AI coaching. It does not articulate substantive positions on confidentiality boundaries, fidelity to source material, handling of distress, distinction between coaching and therapy, treatment of vulnerable populations, and many other dimensions that are essential to ethical coaching practice. These dimensions live in federation adaptations.
This is intentional. The framework starts minimal because the founder does not claim to know in advance what is universal across all coaching traditions. Claims of universality must be earned through observation, not asserted.
3. The inductive posture
CTF grows by convergence. When several federation adaptations independently include a similar requirement, that requirement becomes a candidate for upstreaming into the generic core. This is the central dynamic of the framework’s evolution and is described in detail in ../GOVERNANCE.md, section 5.
4. Relation to federation adaptations
Federation adaptations (CTF-ICF, CTF-EMCC, etc.) are the natural carriers of substantive ethical content. They live in ../declinations/. They may be authored by the federation itself (federation pathway) or by a community member with federation validation (community pathway).
A platform implementing CTF without yet having access to a relevant federation adaptation implements the generic core only. This is a valid form of conformance, with the understanding that it is structurally lighter than implementing a federation-validated adaptation.
Part II — Operational principles (the minimal core)
[Content to be integrated from the founder’s working files. This part articulates the small set of principles that any AI coaching agent must satisfy, regardless of tradition. These are the principles whose violation would make any conformance claim hollow.]
5. Sollicited tool, not conversational companion
[To be integrated.]
6. Non-anthropomorphization
[To be integrated.]
7. Non-substitution to human professionals
[To be integrated.]
8. Auditability by construction
[To be integrated.]
9. Anonymization by construction of audit material
[To be integrated.]
10. Hierarchy of principles in case of conflict
[To be integrated.]
Part III — The quality cycle (operational requirements for platforms)
[Content to be integrated. This part defines the quality engineering practices that a platform must implement to be considered CTF-conformant. The practices are operationalized through the helper agents in ../tools/.]
11. Use of CTF Designer at design phase
[To be integrated.]
12. Use of CTF Test Plan and CTF Robustness before deployment
[To be integrated.]
13. Use of CTF Audit during deployment, on documented periodicity
[To be integrated.]
14. Tripartite supervision (eggspert + certified supervisor + CTF Audit)
[To be integrated.]
15. Material maintenance (design documents, test plans, audit reports)
[To be integrated.]
Part IV — Conformance and certification
16. Self-declaration: minimal commitments
A platform may publicly self-declare its conformance to the generic CTF following the procedure in ../certification-scheme/self-declaration.md. The minimal commitments are summarized there.
17. Adaptation pathway: relationship to federation adaptations
When a federation adaptation exists for a sphere in which the platform operates, the platform may pursue conformance to the adaptation rather than to the generic core. The adaptation includes everything in the generic core, plus federation-specific requirements. See ../certification-scheme/federation-pathway.md and ../certification-scheme/community-pathway.md.
18. Loss of conformance
A platform that substantively departs from its declared commitments must either return to conformance promptly or update its public declaration. Maintaining a declaration that no longer reflects practice is misrepresentation.
Part V — Vocabulary
19. Defined terms
The full bilingual glossary of CTF terms is in ../explainer/glossary.md. The normative document uses these terms in the senses defined there.
20. Rejected terms
CTF refuses certain terms because they carry connotations incompatible with the framework’s posture. The list of rejected terms and the reasons for each rejection will be integrated from the founder’s working files.
Part VI — Revision rules
21. Versioning
The normative document follows the versioning rules described in ../GOVERNANCE.md, section 4.
22. Founding principle as upper boundary
The founding principle inscribed at the head of this document and at the head of GOVERNANCE.md is the upper boundary that all subsequent revisions must respect. No revision may contradict it.
23. Upstreaming from adaptations into the generic core
The mechanism by which substantive content moves from federation adaptations into the generic core is the upstream proposal process described in ../GOVERNANCE.md, section 5. An upstream proposal requires, at minimum:
- Observation of the candidate requirement in at least two independently authored federation adaptations.
- A formal proposal using the proposal template, with the “Upstream” type.
- An opinion from the scientific panel on the quality of the convergence claim (when the panel exists).
- An opinion from the advisory circle on the legitimacy of the convergence (when the circle exists).
- The caretaker’s documented decision.
End of normative document, version 1.0.