Quality Fabric Craft

name: quality-fabric-craft description: Use when authoring or shipping any quality-bound artifact in this AKOS workspace (audience-tagged, closure UAT, sibling-repo work, brand-touching surface, Quality Fabric specialty canonical). Codifies the craft for resolving + composing the 5 axes (audience × channel × scenario × brand × governance) multiplicatively. Triggers on Quality Fabric, compose_UAT, compose_render, 5-axis composition, audience resolution, channel resolution, scenario resolution, brand register, governance binding, multiplicative-AND composition. Pairs with .cursor/rules/akos-quality-fabric.mdc (the WHEN); this skill is the HOW. version: 1.0.0 ratifying_decisions: D-IH-86-CT; D-IH-86-AU authored: 2026-05-24

Quality-Fabric Craft

Codified at I86 Wave R close (drain7) to operationalise the 5-axis composition that the operator ratified at D-IH-86-AU (2026-05-20). The craft turns the multiplicative-AND rule from doctrinal claim into authoring practice. Ratified by D-IH-86-CT alongside the parent rule.

When to use this skill

Read this skill before authoring or revising any quality-bound artifact — anything carrying an audience: frontmatter tag, any closure UAT, any sibling-repo deploy-triggering work, any brand-touching surface (deck slide body, dossier prose, web page, email body), or any Quality Fabric specialty canonical.

This skill assumes you have already read akos-quality-fabric.mdc (the when) + HOLISTIKA_QUALITY_FABRIC.md (the parent meta-doctrine).

Core principles

Principle 1 — Resolve all 5 axes BEFORE drafting prose

The most common failure mode is to draft the artifact first and reverse-engineer the bar after-the-fact. The bar then bends to fit the draft, not vice versa. Always:

1. Audience — FK-resolve audience: against AUDIENCE_REGISTRY.csv. Multi-audience → strictest constraint per downstream axis.
2. Channel — FK-resolve channel: against CHANNEL_TOUCHPOINT_REGISTRY.csv (optional today; advisory). Absence → audience-class default.
3. Scenario — FK-resolve scenario: against PERSONA_SCENARIO_REGISTRY.csv (optional today). Absence → parent-initiative scope description.
4. Brand — always non-null; the variable is the register (CORPINT-internal vs translated-external per akos-brand-baseline-reality.mdc).
5. Governance — always non-null; determined by asset class per PRECEDENCE.md.

Cite the resolved axes in the artifact's frontmatter so the resolution is auditable.

Principle 2 — Multiplicative-AND, not additive, not last-axis-wins

When constraints from two axes appear in conflict, BOTH apply. Don't drop one. Don't average. Don't let the latest-considered axis win. The compose() function is a logical AND across all 5 axes; the artifact must satisfy every constraint from every axis.

When constraints genuinely conflict (e.g., audience demands brevity but governance demands completeness), surface as inline-ratify gate per akos-inline-ratification.mdc — let operator pick the reconciliation rather than silently choosing.

Principle 3 — Per-axis doctrine is the source of truth (don't infer from prior outputs)

Don't infer brand voice from prior outputs you've authored; read BRAND_DO_DONT.md. Don't infer channel norms from intuition; read the per-channel doctrine (when minted) or call out the gap explicitly in notes: frontmatter. Don't infer audience expectations from "what investors want"; read AUDIENCE_REGISTRY.csv row for the specific J-IN class.

Each axis has a canonical SSOT; the skill is to consult the SSOT not to wing it.

Principle 4 — Under-specified artifacts cannot reach PASS

If any of the 5 axes is unresolvable (e.g., audience tag missing; can't determine the right strictness), the artifact is under-specified and cannot reach a PASS verdict in its UAT/UX bar. Either:

• Resolve the axis (via inline-ratify per Principle 2 when contested).
• Carry the under-specification flag explicitly in notes: frontmatter with a closure-target.

Refusing to author under-specified artifacts is correct behaviour, not blocking behaviour.

Principle 5 — Specialty canonicals inherit the fabric

When authoring a new specialty canonical that materialises a compose() function (per parent rule RULE 7), the canonical:

1. Cross-references HOLISTIKA_QUALITY_FABRIC.md as parent meta-doctrine.
2. Names its compose() function explicitly (e.g., compose_UAT(audience, channel, scenario, brand, governance) → 7-class UAT shape).
3. Lands at status: charter if not yet exercised; promotes to status: active per QF §10 criteria.
4. Cites internal precedents per akos-applied-research-discipline.mdc RULE 1.
5. Cites at least one external research source if novel per RULE 2.

Per-axis craft

Audience axis

• Resolve audience: against AUDIENCE_REGISTRY.csv (J-OP / J-AIC / J-IN / J-CU / J-PT / J-AD / J-ENISA / J-RC / J-CO).
• Multi-audience tags use ;-separated list; resolution picks the STRICTEST constraint per downstream axis.
• Read the row's register_class + delivery_format + voice_register columns to derive per-axis constraints.

Channel axis

• Optional today (advisory per akos-external-render-discipline.mdc RULE 7).
• When present, resolve channel: against CHANNEL_TOUCHPOINT_REGISTRY.csv (CHAN-EMAIL-OUTBOUND / CHAN-LINKEDIN-DM / CHAN-WEB-FORM / etc.).
• When absent, fall back to audience-class default channel (look up in AUDIENCE_REGISTRY.csv default_channel column).

Scenario axis

• Optional today.
• When present, resolve scenario: against PERSONA_SCENARIO_REGISTRY.csv.
• When absent, scenario falls back to the artifact's parent-initiative scope description.

Brand axis

• Always non-null.
• Variable: CORPINT-internal register (audience J-OP / J-AIC only) vs translated-external register (any other audience).
• Read BRAND_BASELINE_REALITY_MATRIX.md §3 translation table; never render internal tokens to external audience.

Governance axis

• Always non-null.
• Determined by asset class per PRECEDENCE.md.
• For canonical CSVs: canonical-CSV gate per akos-baseline-governance.mdc.
• For deploy-triggering work: per-vendor gate per akos-deploy-health.mdc.
• For external-delivery surfaces: render-trail per akos-external-render-discipline.mdc.

Pre-flight checklist

1. All 5 axes resolved BEFORE drafting any prose.
2. Frontmatter declares the resolved axes (audience:, channel: if known, scenario: if known, linked_decisions: for governance, brand register implicit in content).
3. Per-axis doctrine consulted (NOT inferred from prior outputs).
4. Multiplicative-AND composition applied; conflicts surfaced as inline-ratify (not silently chosen).
5. Under-specified axes either resolved OR explicitly flagged in notes: with closure-target.
6. If new specialty: cross-references QF parent + names compose() function + cites precedents.
7. UAT report's findings table addresses each axis (PASS/SKIP/N/A per axis).

Anti-patterns

• AP1 — Reverse-engineer the bar from the draft. The draft bends governance to fit; bad doctrine.
• AP2 — Drop the conflicting axis. Multiplicative-AND violated; one stakeholder loses.
• AP3 — Last-axis-wins. Whoever the agent thought of last wins; arbitrary.
• AP4 — Silent reconciliation. Agent picks reconciliation without surfacing; operator can't audit.
• AP5 — Infer brand voice from prior outputs. Doctrinal drift compounds across artifacts; the SSOT loses authority.

Cross-references

• Parent rule: akos-quality-fabric.mdc.
• Parent canonical: HOLISTIKA_QUALITY_FABRIC.md.
• Specialty SSOTs: UAT_DISCIPLINE.md, UX_DISCIPLINE.md, INTER_WAVE_REGRESSION_DISCIPLINE.md, INDEX_INTEGRITY_DISCIPLINE.md, DATAOPS_DISCIPLINE.md, MKTOPS_DISCIPLINE.md, TECHOPS_DISCIPLINE.md.
• Axis SSOTs: AUDIENCE_REGISTRY.csv, CHANNEL_TOUCHPOINT_REGISTRY.csv, PERSONA_SCENARIO_REGISTRY.csv, BRAND_BASELINE_REALITY_MATRIX.md, PRECEDENCE.md.
• Sister skills: inline-ratify-craft, applied-research-craft, external-render-craft.
• Ratifying decisions: D-IH-86-CT (this skill mint), D-IH-86-AU (parent rule ratification).

Source: FraysaXII/openclaw-akos — distributed by TomeVault.