Resources

Resources — Doctrine

The doctrinal corpus: whitepaper, 4 principles, 4 autonomy levels, 6 maturity dimensions, DDAO role, pedagogical manual (9 parts) and deck. The skeleton of the ACF® standard.

iNote

Doctrinal resources are the root of everything: the cards reference them, the regulator guides map onto them, and REASON tools invoke them in their rationale. If something stays unclear, you are likely reading the wrong layer — come back here.

What this contains

Fifteen canonical resources in total: three narrative documents (whitepaper, deck, manual), fourteen structured bricks (principles, levels, dimensions, DDAO). The manual additionally breaks down into a TOC plus nine individually addressable sections.

Exhaustive list

Root documents

acf://whitepaperMarkdownFR + EN

ACF® Whitepaper

The founding doctrinal document. Served as a single Markdown (V1.0 flat PDF extract; no section anchors).

acf://framework/ddaoJSONFR + EN

DDAO — Delegated Decision Agent Officer

Canonical definition of the ACF® governance role responsible for autonomous-agent guardianship.

acf://deckMarkdownFR only

ACF® Slide deck

Presentation deck for the standard — V1.0 ships the French version; English will follow.

Founding principles ×4

See also — READ tool acf.search to search a principle by keyword.

acf://framework/principle/P1JSONFR + EN

P1 — Decision Sovereignty

The organisation retains ultimate responsibility for every agentic decision, whatever the autonomy level.

acf://framework/principle/P2JSONFR + EN

P2 — Doctrinal Traceability

Every automated decision must be reconstructible by pointing to the doctrine, rules and data that produced it.

acf://framework/principle/P3JSONFR + EN

P3 — Ultimate Human Control

Every autonomous agent ships with a human-takeover mechanism, activatable at any time and calibrated to criticality.

acf://framework/principle/P4JSONFR + EN

P4 — Proportional Governance

Control intensity is proportional to the agent’s real autonomy and to the criticality of the decisions it takes.

Autonomy levels ×4

See also — REASON tool acf.assess-autonomy for an agent to recommend an N0–N3 level from a use case.

acf://framework/autonomy-level/N0JSONFR + EN

N0 — Assistance

The agent suggests; the human decides systematically. The human is the kill switch.

acf://framework/autonomy-level/N1JSONFR + EN

N1 — Supervised recommendation

The agent proposes a justified decision. The human validates or rejects on an already-shaped option.

acf://framework/autonomy-level/N2JSONFR + EN

N2 — Conditional execution

Bounded autonomous execution. Above mandate thresholds, escalation is mandatory.

acf://framework/autonomy-level/N3JSONFR + EN

N3 — Autonomous execution

Full autonomous execution within mandate; human control remains a posteriori, kill switch stays active.

Maturity dimensions ×6

See also — REASON tool acf.identify-governance-gaps for a gap analysis mapped onto the six dimensions.

acf://framework/dimension/D1JSONFR + EN

D1 — Strategy & governance

Alignment of agentic AI with business strategy, dedicated governance, executive sponsorship.

acf://framework/dimension/D2JSONFR + EN

D2 — Doctrine & framework

Adoption of a reference framework, published internal doctrine, team training.

acf://framework/dimension/D3JSONFR + EN

D3 — Design & technical control

Secure agent design — mandates, kill switches, observability, foundation-model mastery.

acf://framework/dimension/D4JSONFR + EN

D4 — Accountability & roles

Explicit role definitions (DDAO, DPO, CISO, Compliance, Business) with clear RACI on agentic decisions.

acf://framework/dimension/D5JSONFR + EN

D5 — Regulatory compliance

Mastery of AI Act, GDPR, DORA, NIS2, ISO 42001 obligations by sector and jurisdiction.

acf://framework/dimension/D6JSONFR + EN

D6 — Audit & continuous improvement

Periodic internal audit, incident review, doctrine update cadence.

Pedagogical manual × 1 + TOC + 9 sections

V1.0 only ships the French manual. A client requesting English (Accept-Language: en) receives the French content with is_fallback: true.

acf://manualMarkdownFR only

ACF® Manual — full

Full pedagogical manual concatenating all 9 parts. Large payload — prefer the TOC + sections for incremental reads.

acf://manual/tocJSONFR only

ACF® Manual — table of contents

Structured TOC of the manual — 9 parts with titles, page ranges and section URIs.

acf://manual/section/{1..9}MarkdownFR only

ACF® Manual — individual section

One of the 9 manual parts. Stable URI; useful for citing a precise passage in an audit reply.

Fetching a doctrine resource

Framework resources return structured JSON (title, summary, doctrine, related_fiches, related_dimensions fields). Narrative resources return Markdown with a served_locale and is_fallback frontmatter at the top.

fetch-principle.tsts

// Read principle P3 — Ultimate Human Control
const { contents } = await client.readResource({
  uri: "acf://framework/principle/P3",
});

const payload = JSON.parse(contents[0].text);
// {
//   "id": "P3",
//   "code": "P3",
//   "title": { "fr": "Contrôle humain ultime", "en": "Ultimate Human Control" },
//   "summary": { "fr": "...", "en": "..." },
//   "doctrine": { "fr": "...", "en": "..." },
//   "related_fiches": ["ACF-07", "ACF-09", "ACF-14"],
//   "related_dimensions": ["D3", "D6"]
// }

// Read the whitepaper in French
const wp = await client.readResource({ uri: "acf://whitepaper" });
console.log(wp.contents[0].mimeType); // "text/markdown"
console.log(wp.contents[0].text);     // ---\nserved_locale: fr\nis_fallback: false\n---\n...