fix(evolution-curve): settle the capability axis as human-assessed

ARCHITECTURE.md

@@ -48,7 +48,7 @@ Every other artifact — audit rules, teardown scores, memory policies, Trust Co
   │  aux-audit                                             │
   │   1. validate spec against agent-spec.schema.yaml      │
   │   2. run heuristic rules (aux-heuristics.yaml)         │
-  │   3. place on trust + evolution axes                   │
+  │   3. place on the trust axis (capability: see #5)      │
   │   4. classify gaps (trust-gap-taxonomy.yaml)           │
   │   5. emit Markdown + JSON + SARIF                      │
   └────────────────────────────────────────────────────────┘

CHANGELOG.md

@@ -33,6 +33,7 @@ Publishing the GitHub release and tagging `v0.2` is a manual step after merge.
 
 ### Notes
 
-- `evolution_stage` is reported as `schema-undefined`. The Evolution Curve is
-  referenced in docs but has no schema; the CLI will not mint `aux.` IDs.
+- `evolution_stage` is reported as `human-assessed` and is always `null`. The
+  capability axis is scored by a human in a teardown, not derived from a spec
+  ([#5](https://github.com/auxfirst/trustkit/issues/5)).
 - A spec with no evaluation transcripts is capped at heuristic level 2.

README.md

@@ -163,7 +163,7 @@ See **[LICENSE](LICENSE)** for the CC BY 4.0 legal code and **[LICENSING.md](LIC
 
 ## Status
 
-**v0.2 · 2026-04-19** — First public release. Onboarding layer live: audience-routed docs (R1–R4) + four forwardable artifacts + two canonical patterns + six first-wave repo READMEs + canonical schemas. `aux-audit` v0.1 is implemented, tested, and wired into CI as a GitHub Action. Next: `npm publish` so `npx aux-audit` resolves without a clone; list the Action on the Marketplace; publish a schema for the Evolution Curve, which `aux-audit` currently reports as `schema-undefined`.
+**v0.2 · 2026-04-19** — First public release. Onboarding layer live: audience-routed docs (R1–R4) + four forwardable artifacts + two canonical patterns + six first-wave repo READMEs + canonical schemas. `aux-audit` v0.1 is implemented, tested, and wired into CI as a GitHub Action. Seven patterns close all 12 named trust gaps, enforced in CI. Next: `npm publish` so `npx aux-audit` resolves without a clone; list the Action on the Marketplace; name the Evolution Curve stages for teardowns ([#5](https://github.com/auxfirst/trustkit/issues/5) — the axis is human-assessed, not something `aux-audit` computes).
 
 ---
 

docs/audit.md

@@ -22,8 +22,9 @@ npx aux-audit run ./agent-spec.yaml
 - **Exit codes:** `0` clean · `1` findings at or above `--fail-on` · `2` invalid spec (nothing graded).
 
 The audit grades the **spec**, not the running product. Level 3 (“robust”) is
-unreachable without evaluation transcripts. `evolution_stage` stays `null` until
-an Evolution Curve schema is published.
+unreachable without evaluation transcripts. `evolution_stage` is always `null`:
+capability is what a product *does*, which a spec cannot state, so it is
+assessed by a human in a teardown ([#5](https://github.com/auxfirst/trustkit/issues/5)).
 
 ## GitHub Action
 

docs/releases/v0.2.md

@@ -19,7 +19,7 @@ TrustKit is now a **runnable standard**, not only a set of YAML files.
 
 ## What it is not yet
 
-- The Evolution Curve has no schema. `evolution_stage` is `null`.
+- The Evolution Curve stage names are still open ([#5](https://github.com/auxfirst/trustkit/issues/5)). `evolution_stage` is `null` in `aux-audit` by design — it belongs to teardowns, not to the CLI.
 - The Action is not listed on the Marketplace until you submit it.
 - `npx aux-audit` without a local checkout needs `npm publish` from
   `packages/aux-audit`.

onboarding/appendix-a-vocabulary.md

@@ -33,5 +33,5 @@ Two tables. One for GitHub words (useful to R1 and R4 if they venture there). On
 | **Memory policy** | Declared scopes, retention, and user controls for persistent memory. | [schemas/memory-policy.schema.yaml](../schemas/memory-policy.schema.yaml) |
 | **Pattern** | A named, reusable design move that closes a specific gap. | [repos/agentic-ux-patterns](../repos/agentic-ux-patterns/README.md) |
 | **Teardown** | A structured review of a shipped product: `analysis.md` + `score.json`. | [repos/agent-ux-teardowns](../repos/agent-ux-teardowns/README.md) |
-| **Evolution curve** | The capability axis, independent of trust. **No schema yet** — `aux-audit` reports `schema-undefined` rather than inventing stages. | referenced from [schemas/trust-architecture.yaml](../schemas/trust-architecture.yaml); tracked as a schema Issue |
+| **Evolution curve** | The capability axis. Assessed by a human in a teardown, never computed from a spec — a spec says what a product claims, capability is what it does. Stage names are still open. | [trustkit#5](https://github.com/auxfirst/trustkit/issues/5); home is [repos/agent-ux-teardowns](../repos/agent-ux-teardowns) |
 | **Definition / Executable / Reference** | The three kinds of repo. Every repo is exactly one. | [REPO_STRUCTURE.md](../REPO_STRUCTURE.md) |

packages/aux-audit/README.md

@@ -35,7 +35,7 @@ Change the YAML in `schemas/`, and the audit changes with it. No heuristic text,
 Two consequences, both deliberate:
 
 - **Level 3 requires evidence.** "Robust" means *works under novel, adversarial, or edge cases* — prose cannot demonstrate that. A spec with no `evaluation.golden_transcripts` or `evaluation.failure_transcripts` is capped at level 2 on every heuristic, however well written. This is what stops the score from being gamed by rewriting guarantees.
-- **`evolution_stage` is `null`.** The AUX docs reference an Evolution Curve as the capability axis, but no schema publishes its stages. Rather than invent IDs in the `aux.` namespace, the field reports `evolution_stage_status: "schema-undefined"` and stays empty until the curve ships. See [issue tracker](https://github.com/auxfirst/trustkit/issues).
+- **`evolution_stage` is always `null` — by design, not pending.** A spec states what a product *claims about itself*; capability is a claim about what it *does*. No schema turns a declaration into an observation, so the Evolution Curve is assessed by a human in [`agent-ux-teardowns`](../../repos/agent-ux-teardowns) and this tool reports `evolution_stage_status: "human-assessed"`. Decided in [trustkit#5](https://github.com/auxfirst/trustkit/issues/5).
 
 ## Install
 

packages/aux-audit/src/report/markdown.ts

@@ -86,7 +86,7 @@ export function toMarkdown(report: AuditReport): string {
   lines.push(
     `<sub>aux-audit ${report.meta.tool_version} · heuristics ${report.meta.heuristics_version} · trust architecture ${report.meta.trust_architecture_version} · taxonomy ${report.meta.taxonomy_version}. ` +
       `Scores describe the **spec**, not the running product: aux-audit can prove a mechanism was never declared, never that a declared one works. ` +
-      `Capability axis (\`evolution_stage\`) is omitted — the Evolution Curve has no published schema. ` +
+      `Capability axis (\`evolution_stage\`) is not scored here by design: a spec states what a product claims, not what it does. ` +
       `[The standard](https://github.com/auxfirst/trustkit) · [auxfirst.com](https://auxfirst.com)</sub>`,
   );
   lines.push("");

packages/aux-audit/src/score.ts

@@ -202,7 +202,7 @@ export function audit(spec: AgentSpec, options: AuditOptions = {}): AuditReport
     grade: grade(score),
     trust_stage,
     evolution_stage: null,
-    evolution_stage_status: "schema-undefined",
+    evolution_stage_status: "human-assessed",
     heuristics,
     trust_stages,
     issues,

packages/aux-audit/src/types.ts

@@ -70,12 +70,17 @@ export interface AuditReport {
   grade: Grade;
   trust_stage: string | null;
   /**
-   * Null until the Evolution Curve is published as a schema. The capability
-   * axis is referenced by the AUX docs but has no canonical stage list, and
-   * aux-audit will not invent IDs in the `aux.` namespace.
+   * Always null, by design — not pending.
+   *
+   * A spec describes what a product *declares about itself*; capability is a
+   * claim about what it actually does. No schema turns a declaration into an
+   * observation, so the Evolution Curve is assessed by a human in
+   * `agent-ux-teardowns` and is out of scope for this tool permanently.
+   *
+   * Decided in auxfirst/trustkit#5 (option B).
    */
   evolution_stage: null;
-  evolution_stage_status: "schema-undefined";
+  evolution_stage_status: "human-assessed";
   heuristics: HeuristicResult[];
   trust_stages: TrustStageResult[];
   issues: Issue[];

packages/aux-audit/test/audit.test.ts

@@ -121,10 +121,13 @@ test("the report is deterministic for the same spec", () => {
   assert.deepEqual(audit(spec, { now }), audit(spec, { now }));
 });
 
-test("evolution_stage stays null while the Evolution Curve has no schema", () => {
-  const report = strong();
-  assert.equal(report.evolution_stage, null);
-  assert.equal(report.evolution_stage_status, "schema-undefined");
+test("evolution_stage is permanently out of scope, not pending", () => {
+  // trustkit#5 option B: capability is assessed by a human in teardowns,
+  // never computed from a spec. The status must not read as future work.
+  for (const report of [strong(), weak()]) {
+    assert.equal(report.evolution_stage, null);
+    assert.equal(report.evolution_stage_status, "human-assessed");
+  }
 });
 
 test("level 3 requires evidence — a spec with no transcripts is capped at present", () => {

repos/agent-ux-teardowns/README.md

@@ -31,7 +31,8 @@ Every `analysis.md`:
 3. **Where the contract breaks** — named failures from [`trust-gap-classifier`](../trust-gap-classifier).
 4. **Heuristic scorecard** — 10 heuristics, scored 0–3, with evidence.
 5. **Trust stage** — placed on the Trust Architecture ladder.
-6. **The fix** — one concrete pattern that would move it up a stage.
+6. **Evolution stage** — placed on the capability axis. This is the *only* place the capability axis is assessed: `aux-audit` reads a spec, and a spec states what a product claims rather than what it does. Stage names are still open — see [trustkit#5](https://github.com/auxfirst/trustkit/issues/5) — so `evolution_stage` stays `null` in teardowns until they land.
+7. **The fix** — one concrete pattern that would move it up a stage.
 
 Every `score.json`:
 
@@ -40,7 +41,7 @@ Every `score.json`:
   "product": "ChatGPT",
   "version": "4o, 2026-04",
   "trust_stage": "contextual",
-  "evolution_stage": "personally-intelligent",
+  "evolution_stage": null,
   "heuristics": {
     "aux.H01": 2, "aux.H02": 1, "aux.H03": 3,
     "aux.H04": 2, "aux.H05": 1, "aux.H06": 2,

repos/aux-audit/README.md

@@ -18,7 +18,7 @@ npx aux-audit run ./agent-spec.yaml
 score: 72
 grade: B
 trust_stage: contextual
-evolution_stage: null          # schema-undefined until the Evolution Curve ships
+evolution_stage: null          # human-assessed in teardowns, never computed here
 issues:
   - id: aux.H01
     type: visibility_of_intent
@@ -41,7 +41,7 @@ recommendations:
 
 1. **Parses** an agent spec (JSON/YAML describing the product: surface, memory model, tool use, autonomy model, failure handling).
 2. **Runs** the rule set from [`aux-frameworks`](../aux-frameworks) against it.
-3. **Scores** trust maturity from the Trust Architecture. Capability (`evolution_stage`) is `null` until an Evolution Curve schema is published.
+3. **Scores** trust maturity from the Trust Architecture. Capability (`evolution_stage`) is always `null` here — it is assessed by a human in [`agent-ux-teardowns`](../agent-ux-teardowns), per [trustkit#5](https://github.com/auxfirst/trustkit/issues/5).
 4. **Emits** a scored report as Markdown, JSON, or SARIF (for CI integration).
 
 ## Why CLI-first

repos/aux-frameworks/README.md

@@ -14,9 +14,6 @@
   /trust-architecture       # 4-stage trust maturity ladder
     schema.yaml
     stages.yaml
-  /aux-evolution-curve      # capability maturity axis
-    schema.yaml
-    stages.yaml
   /aux-audit                # audit rules derived from the above
     schema.yaml
     rules.yaml