openmerit 0.1.3 → 0.1.5

package/CHANGELOG.md

@@ -0,0 +1,26 @@
+# Changelog
+
+## 0.1.5 — 2026-09-23
+
+OpenMerit now ships as one npm package with a Pi extension, a reusable
+coordinator, and a harness-neutral protocol.
+
+### Changed
+
+- Pi issues typed, bounded work to the coding harness and checks structured
+  results and durable evidence before advancing the model-improvement lifecycle.
+- Project state, audit events, and evidence manifests live under `.openmerit/`
+  in the product repository.
+- Baseline assessment, candidate trials, frontier verification, supervised
+  swaps, post-swap verification, and rollback follow a confirmed project policy.
+- The package exposes `openmerit/core`, `openmerit/protocol`, and `openmerit/pi`.
+- Node.js 22.19 or newer and Pi 0.87 are required for the Pi integration.
+
+### Removed from the 0.1.4 package
+
+- The background watcher, direct provider clients, standalone CLI, and
+  session-bound route comparison flow.
+- Automatic use of the earlier home-directory state. The new project-local
+  state starts with a confirmed task profile and leaves older files untouched.
+
+See [the README](README.md) for installation and the current command surface.

package/README.md

@@ -1,342 +1,113 @@
-# openmerit
+# OpenMerit
 
-An **experimental model-merit harness** for [pi](https://pi.dev). Start
-pi with model A, and the OpenMerit extension observes
-its completed trace, trials B and C sequentially through pi, compares quality,
-price, and latency, and recommends a model for that exact pi session. The
-shipped policy is supervised: you review and apply a recommendation with
-`/openmerit apply`. Automatic swaps are available as an explicit opt-in.
+OpenMerit is a harness-neutral orchestration layer that helps a coding harness continuously find and maintain the Pareto frontier of models for a real product task.
 
-```text
-pi task A → saved pi trace → extension trial job → pi trials B, C → scored frontier
-    ↑                                                     │
-    └──── pi.setModel() ← approval/policy ← recommendation ┘
-```
-
-## How it maps to the design
-
-1. **Instruction file + policy file** — [`instructions/OPENMERIT.md`](instructions/OPENMERIT.md)
-   goes into the main harness's context (AGENTS.md-style);
-   [`instructions/openmerit.policy.json`](instructions/openmerit.policy.json)
-   (copied to `~/.openmerit/policy.json` by `openmerit init`) gates autonomy:
-   supervised recommendations by default, optional auto-apply thresholds,
-   budgets, and provider allow-lists. Public-benchmark relevance lives in
-   [`src/benchmarks.ts`](src/benchmarks.ts) (task category → benchmarks that
-   matter) plus a refreshable digest at `~/.openmerit/benchmarks/digest.json`.
-2. **Trial workflow** — the pi extension starts a session-bound comparison
-   after a supported task settles. It runs candidate trials sequentially under
-   configured candidate-trial caps. `openmerit watch` remains an optional CLI
-   mode. Pi's traces and JSON event stream supply results, cost,
-   latency, and errors; a judge or known-ground-truth scorer measures quality.
-3. **Per-task pareto frontier** — trials are keyed by task signature
-   ([`src/frontier.ts`](src/frontier.ts), 3 objectives: quality ↑, price ↓,
-   latency ↓). The aggregate across all of an agent's tasks is the union of
-   its task frontiers (`openmerit frontier`).
-4. **Model discovery** — the active Pi session's scoped models (or Pi's
-   authenticated available-model registry when unscoped) are the candidate
-   pool. OpenRouter is an optional route and enrichment source: its catalog can
-   be snapshotted and its benchmark signals can improve the shortlist
-   ([`src/catalog.ts`](src/catalog.ts) + [`src/daemon.ts`](src/daemon.ts)).
-5. **Informing the main harness** — recommendations land in
-   `~/.openmerit/recommendations.jsonl`. The pi extension
-   ([`extension/openmerit.ts`](extension/openmerit.ts)) reports progress, polls it, and applies
-   swaps with `pi.setModel()` — automatically when the policy gate passes,
-   otherwise after user approval (`/openmerit`, `/openmerit apply`). The same
-   recommendation carries the updated fallback, used when the active model
-   starts failing.
-
-## Install from npm
-
-You need Node 22.18+, pi 0.85.1+, and at least two eligible model routes
-authenticated in Pi. OpenRouter is supported but not required. Install the
-package into pi, then initialize its policy and state:
-
-```bash
-pi install npm:openmerit
-npx --yes openmerit init
-npx --yes openmerit verify
-pi list
-```
-
-`pi install` makes the extension and its trial engine available to pi. The
-one-off `npx` command creates `~/.openmerit/policy.json`; install OpenMerit
-globally with `npm install --global openmerit` only if you also want persistent
-shell access to `openmerit status`, `doctor`, `verify`, `frontier`, or the optional watcher. Install
-the extension from only one source—remove any older copied `openmerit.ts` first
-so pi does not load it twice.
-
-For development from a local checkout instead, run:
-
-```bash
-npm ci
-node dist/cli.js init
-pi install "$PWD"
-```
-
-Keep that checkout available because pi loads a local package from its path.
-
-## Run OpenMerit with pi
-
-1. Initialize OpenMerit if you did not do so during installation:
-
-   ```bash
-   npx --yes openmerit init
-   ```
-
-   `init` creates `~/.openmerit/policy.json` if missing and **keeps an existing
-   policy**. The shipped policy uses `"mode": "recommend"` and does not change
-   the active model without approval.
+OpenMerit does not perform the intelligent work. It nudges a connected coding harness, such as Pi, with typed outcome requests. The harness understands the project, establishes evaluations and observability, gathers evidence, discovers and tests candidates, calculates the Pareto frontier, and performs an authorized model change. OpenMerit validates the returned contracts, preserves evidence and preferences, and advances the approved lifecycle.
 
-2. Authenticate the providers you want to compare in Pi, using `/login` or
-   Pi's normal environment/model configuration. OpenMerit takes its eligible
-   model routes, capabilities, prices, and credentials from Pi; judge and
-   strategist calls use the same routes and do not require duplicate keys.
+The npm package `openmerit` contains the Pi extension, the reusable coordinator,
+and the protocol in one install. The Pi adapter is the first integration; other
+harnesses can use the same core and protocol exports.
 
-   For example, verify one provider without printing its credential:
+## Why OpenMerit
 
-   ```bash
-   pi auth check --provider openai --json
-   ```
+The cheapest model, fastest model, and highest-quality model are often different. Public benchmarks are useful before product evidence exists, but they cannot establish which model is best for a particular workload. OpenMerit connects the initial choice to production evidence and repeated reassessment.
 
-   Add an OpenRouter route the same way if you want its routed catalog:
+It distinguishes:
 
-   ```bash
-   pi auth check --provider openrouter --json
-   ```
+- one-run observations from rate, percentile, consistency, and reliability claims;
+- missing evidence from poor performance;
+- feasible candidates from dominated, unresolved, or ineligible candidates;
+- an automatic nudge from an intelligent action performed by the harness;
+- a recommendation from an authorized, verified model swap.
 
-   An `OPENROUTER_API_KEY` in the process environment or
-   `~/.openmerit/.env` additionally enables OpenRouter catalog and public-
-   benchmark enrichment. It is optional for native-provider comparisons. If
-   using the file, protect it:
+## Current capabilities
 
-   ```bash
-   nano ~/.openmerit/.env
-   chmod 600 ~/.openmerit/.env
-   ```
+- Pi 0.87 extension with a single `/openmerit` command surface.
+- Confirmed task profiles, objectives, constraints, sample requirements, and evaluation budgets.
+- The complete initial quality, reliability, cost, latency, efficiency, and human-intervention metric catalogue.
+- Baseline-first evidence collection followed by budgeted candidate discovery and controlled challenger trials.
+- A normative, uncertainty-aware Pareto definition owned by OpenMerit.
+- Harness-calculated frontier results independently checked by OpenMerit's conformance verifier.
+- Versioned result schemas for all ten intents, reusable across harness adapters and exposed dynamically by Pi.
+- Supervised graduation: early swaps require approval, every swap requires verification, and regressions can trigger rollback.
+- Project-local snapshots, redacted append-only audit events, evidence manifests, and optional best-effort event exporters under `.openmerit/`.
+- Model-catalog change detection that can nudge the harness to reassess a mature baseline.
+- User-confirmed task-count, elapsed-time, regression, catalogue-change, and post-swap check policies.
+- Idempotent harness signals, durable pending work, native wakeup planning, and explicit scheduler capability gaps.
 
-   Pi does not read OpenMerit's `.env` file for its ordinary sessions, so an
-   OpenRouter route still needs Pi authentication. Older `model_search/.env`
-   files are not read.
+## Responsibility boundary
 
-3. Verify the extension appears in `pi list`. The instruction file
-   [`instructions/OPENMERIT.md`](instructions/OPENMERIT.md) can be added to a
-   pi project's AGENTS.md for agent context, but the extension does
-   not require it. Run `npx openmerit verify` for a provider-free core self-test,
-   then `npx openmerit doctor` after starting Pi once to inspect the eligible
-   route snapshot and configuration without printing secrets.
+| OpenMerit | Coding harness |
+| --- | --- |
+| Defines typed outcomes and Pareto conformance rules | Understands the product and user task |
+| Tracks lifecycle, policy, budget, and evidence references | Builds evals and observability |
+| Determines which approved nudge is due | Collects and calculates metrics |
+| Verifies returned structure and frontier conformance | Discovers candidates and calculates the frontier |
+| Enforces confirmation and rollback policy | Applies and verifies an authorized model change |
 
-4. Start pi in one terminal with any configured model. For example:
+Jev System One, when configured, belongs to the coding harness. Pi may use it aggressively for bounded ranking, triage, and fast judgment. OpenMerit contains no Jev client and accepts no Jev credential.
 
-   ```bash
-   pi --provider openrouter --model openai/gpt-4o-mini
-   ```
+## Install with Pi
 
-   For a first text task, ask: “Give the shortest valid word ladder from cat
-   to dog. Each step changes one letter and must be a common English word.
-   Return only the path.” Wait for A to finish and leave the pi session open.
-   The extension automatically starts B and C **sequentially through their
-   selected Pi routes**,
-   reports each score in Pi, and writes a recommendation for this exact
-   session. With the shipped supervised policy, use
-   `/openmerit` inside pi to inspect the evidence and `/openmerit apply` to
-   switch. Then send a second message to see which model actually handles it.
-   You do not need a watcher terminal or the standalone `trial` command.
+Requirements:
 
-   Candidate comparisons have no tools by default, even when the observed task
-   used tools. See **Alpha boundaries** before explicitly enabling candidate
-   tools.
+- Node.js 22.19 or newer
+- Pi 0.87 with a supported model provider configured
 
-   To opt in to automatic swaps, edit `~/.openmerit/policy.json`, set
-   `"mode": "auto"` and `"auto_apply.enabled": true`, review the score-gain
-   and price-ratio thresholds, and run the next task.
-
-### Try an invoice-to-JSON task
-
-Use the same one-terminal setup. Start pi with a vision-capable model, for
-example `openai/gpt-4o-mini`. In pi, type `@` to select
-[`benchmark/invoice_ocr/data/invoice_01_row_2.jpg`](benchmark/invoice_ocr/data/invoice_01_row_2.jpg)
-and paste the **entire, unchanged** text from
-[`examples/invoice-prompt.txt`](examples/invoice-prompt.txt) into the same
-message. Pi also accepts pasted or dragged images. Wait for the JSON answer;
-keep pi open while the extension trials two other vision models. OpenMerit uses
-the existing benchmark's visibility-audited exact-field scorer when the saved
-task contains the exact example prompt and original bytes of a pinned JPEG.
-Pi may resize an attached image or add a file header to the prompt; in that
-case, the run uses the vision judge that sees the image and answer. Other
-uploaded invoices also use that judge because they have no ground truth.
-
-A real pinned-invoice run produced this trial output (models and scores will
-vary between runs):
-
-```text
-[openmerit] task 0e7642522dc5: A=openai/gpt-4o-mini score=1.00 from pi trace
-[openmerit] task 0e7642522dc5: google/gemma-3-12b-it score=0.31 cost=$0.0001
-[openmerit] task 0e7642522dc5: mistralai/ministral-14b-2512 score=1.00 cost=$0.0007
-[openmerit] task 0e7642522dc5: selected mistralai/ministral-14b-2512; auto=false
-```
-
-Pi candidate runs rely on the explicit invoice prompt for JSON structure.
-The standalone invoice benchmark sends a strict OpenRouter `response_format`
-schema, so its published scores are not directly comparable to these pi runs.
-
-### Inspect or troubleshoot a run
-
-`npx openmerit doctor` checks Pi, policy, eligible routes and prices, saved
-session state, duplicate package sources, append-only files, and optional
-OpenRouter enrichment. Add `--json` for a sanitized diagnostic report suitable
-for a bug report; it contains no credentials, prompts, or trace contents.
-`npx openmerit verify` runs an offline self-test of atomic state, JSONL recovery,
-route preservation, policy evidence, and neutral events without contacting a
-provider. `npx openmerit status` shows the latest pi model and pending recommendations;
-`npx openmerit frontier` shows measured quality, blended price, latency, and
-the chosen frontier per task. `/openmerit` inside pi shows the current model,
-fallback, the model currently being compared, completed models with quality,
-cost, and latency, trial budget, pending recommendations, and any exact-task result
-measured in another session. When the gate declines an automatic swap, it
-prints reasons and `/openmerit apply` remains available.
-
-The daily trial-count and dollar limits come from `~/.openmerit/policy.json`.
-If a limit is reached, the extension reports why it skipped the comparison;
-the count resets at midnight UTC. You may raise `budgets.max_trials_per_day`
-for local experiments while keeping `budgets.max_usd_per_day` as a conservative
-candidate-spend threshold.
-
-If no comparison starts after Pi settles, confirm that pi loaded the extension
-(`pi list`), the task finished, and the pi session is saved (do not use
-`--no-session`). Image candidates must advertise image input in Pi's model
-registry. The extension queues
-completed tasks from its current session and runs one comparison at a time.
-Closing or switching the session cancels the active job.
-Candidate runs use the same text and uploaded image bytes, but they do not
-replay earlier answers or file changes. An exact task in another session
-(including identical image bytes) appears as **advice**, not a pending swap;
-the new session still gets its own comparison.
-
-The optional `trial` command runs controlled text-task comparisons through the
-same exact Pi provider routes and credential store as the automatic session
-path:
-
-```bash
-npx openmerit trial examples/task.example.json --rounds 3
+```sh
+pi install npm:openmerit
+pi list
 ```
 
-`initial_model` is the stable `vendor/model` identity. When Pi exposes that
-model through more than one provider, set `initial_route` to
-`provider:model-id` (for example `openai:gpt-4o-mini` or
-`openrouter:openai/gpt-4o-mini`). Judge and strategist calls also run through
-Pi. An OpenRouter key only adds optional catalog and public-benchmark metadata.
+When Pi enters an unconfigured product with an interactive UI, OpenMerit automatically issues the setup nudge. Pi infers a task profile, asks the user to confirm requirements and an evaluation budget, establishes task-specific eval and observability artifacts, verifies them, and reports evidence to OpenMerit. `/openmerit setup` remains available as an explicit recovery or rerun command.
 
-## Alpha boundaries
+Run `/openmerit doctor` inside Pi to check the installation. OpenMerit keeps
+project evidence under `.openmerit/` in the product repository. Version 0.1.5
+replaces the 0.1.4 background watcher and standalone CLI with this
+harness-driven lifecycle; it does not migrate the earlier home-directory state.
 
-- With the extension installed and at least two eligible Pi routes, each
-  supported settled task can start comparison calls automatically. Candidate,
-  judge, and strategist requests send the task text, attached files or images,
-  and candidate output to the configured providers and can incur charges. Use
-  non-sensitive test tasks and conservative account limits while evaluating
-  this alpha.
-- The extension queues tasks from its current pi session and compares one at a
-  time. Candidate runs use isolated temporary copies of the original working
-  directory and have no Pi tools by default. Their temporary changes are
-  discarded and recorded as counts, and raw Pi JSON event streams are saved
-  under `~/.openmerit/traces/trials/`.
-  Candidate sessions replay the task text and images, not earlier conversation
-  context or workspace changes.
-- Set `OPENMERIT_PI_TRIAL_TOOLS` to an explicit comma-separated allowlist such
-  as `read,grep,find,ls` to let candidate models use tools; `none` keeps them
-  disabled. Any tool access is an advanced opt-in: the copied working directory
-  prevents ordinary project writes from touching the original, but it is not an
-  OS sandbox. Tools may accept absolute paths, and shell tools may access the
-  network. Keep the no-tools default for untrusted or sensitive projects.
-- File attachments that Pi records as `<file name="…">` are copied into every
-  candidate sandbox and passed back to Pi as `@` file inputs. This covers PDFs,
-  CSVs, spreadsheets, and other files that Pi can open; OpenMerit does not
-  implement a separate parser for them.
-- `ledger.json` counts reported candidate, rubric, judge, and strategist spend
-  for automatic session comparisons plus the daily candidate count.
-  `max_usd_per_trial` is a conservative admission estimate based on known Pi
-  prices and a 4K answer; it is not a provider-side hard cap. Routes without
-  known pricing are excluded from automatic comparisons and cannot auto-apply.
-- Each comparison has one observed baseline plus a small candidate slate and
-  one quality score per answer. Treat recommendations as experimental evidence,
-  not a universal model ranking.
-- Candidate execution, judging, and strategy use the provider/model routes
-  exposed by Pi. OpenRouter remains an optional route plus catalog/benchmark
-  enrichment source. `HarnessAdapter`, `ModelProviderAdapter`,
-  `ObservationSource`, and `EventSink` remain separate integration boundaries;
-  Pi and local JSONL are the implementations shipped in this release.
-- Candidate subprocesses can use Pi built-ins and custom/local routes available
-  without loading extensions (for example routes from Pi's model
-  configuration). A provider registered only at runtime by another extension
-  is visible in the route snapshot but cannot yet be executed by the isolated
-  subprocess.
-- JSON state snapshots are replaced atomically. Append-only readers skip and
-  report malformed or interrupted lines while retaining later valid records.
-  A job owned by a crashed process is reclaimable instead of remaining stuck
-  in `running`; completed jobs remain final.
+## Commands
 
-## State layout (`~/.openmerit/`)
+- `/openmerit` or `/openmerit status` — show evidence readiness and lifecycle stage.
+- `/openmerit setup` — explicitly rerun or recover automatic setup.
+- `/openmerit assess` — request an evidence-sufficiency assessment now.
+- `/openmerit logs` — show the canonical audit, exporter-error, evidence, and state paths.
+- `/openmerit frontier` — ask the harness to calculate a frontier now; OpenMerit verifies it.
+- `/openmerit approve` — approve a verified proposal during supervised graduation.
+- `/openmerit pause` and `/openmerit resume` — control proactive nudges.
+- `/openmerit doctor` — show sanitized adapter diagnostics.
 
-| file | contents |
-|---|---|
-| `policy.json` | the policy file (gate thresholds, budgets, intervals) |
-| `harness-state.json` | current/fallback routes, Pi's eligible route snapshot, latest session and settled task |
-| `recommendations.jsonl` | append-only session-bound recommendations, routes, evidence, gate reasons, and status updates |
-| `trials.jsonl` | every model trial point, including its provider route when known |
-| `traces/observations.jsonl` | task observations extracted from session traces |
-| `events.jsonl` | versioned provider-neutral observation, trial, and recommendation events for future sinks |
-| `traces/trials/*.jsonl` | raw Pi JSON event streams for candidate trials |
-| `catalog/snapshot.json` + `candidates.json` | catalog snapshot (including input modalities) + new-model queue |
-| `benchmarks/digest.json` | public-benchmark scores per model (seed + refresh) |
-| `ledger.json` | daily trial spend (budget enforcement) |
-| `watch/processed.json` | latest task handled by optional CLI watcher |
-| `watch/jobs/*.json` | per-session comparison status and retry marker |
+## Documentation
 
-## Benchmarks
+- [Getting started](docs/getting-started.md)
+- [Architecture and responsibility boundary](docs/architecture.md)
+- [Metrics and evidence](docs/metrics-and-evidence.md)
+- [Improvement lifecycle](docs/lifecycle.md)
+- [Harness-neutral automation](docs/automation.md)
+- [Pi extension guide](docs/pi-extension.md)
+- [Building another harness adapter](docs/adapter-guide.md)
+- [Pareto conformance specification](docs/pareto-spec.md)
+- [Project state and operations](docs/operations.md)
+- [Security](docs/security.md)
+- [Testing](docs/testing.md)
+- [Current limitations and roadmap](docs/roadmap.md)
 
-From a source checkout, the benchmark runners are TypeScript and execute
-directly on Node 22.18+; no transpilation step or Python environment is
-required. The slim npm artifact keeps only the pinned invoice data needed by
-runtime scoring; use the repository checkout for these developer commands:
+## Repository layout
 
-```bash
-npm run benchmark:invoice -- --models google/gemini-3.8-flash
-npm run benchmark:invoice:round2 -- --verify-only
-npm run benchmark:text-to-sql -- --verify-only
-npm run benchmark:policy:validate
-npm run benchmark:policy -- --verify-only
-npm run benchmark:verify-recorded
-```
+- `packages/protocol` — harness-neutral types for intents, results, metrics, policies, and frontier evidence.
+- `packages/core` — adapter SDK, durable coordinator, pluggable persistence, policy, and frontier conformance verification.
+- `packages/pi` — Pi extension that turns OpenMerit intents into harness work.
+- `docs` — user, operator, architecture, and normative documentation.
 
-- `benchmark/invoice_ocr/` — two real vision/OCR model slates over the same
-  pinned 10-invoice dataset.
-- `benchmark/text_to_sql/` — executable SQLite evaluation over 10 analytical
-  tasks.
-- `benchmark/policy_adjudication/` — 12 adversarial reimbursement-policy cases,
-  including prompt-injection cases and frozen input hashes.
+The workspace modules are private development boundaries. The published
+package exposes `openmerit/core`, `openmerit/protocol`, and `openmerit/pi`.
 
-All three benchmark families preserve raw JSONL observations, audited summaries,
-cost/latency/quality rankings, and Pareto layers. `npm run typecheck` covers both
-the application and every benchmark runner. `npm run benchmark:verify-recorded`
-replays the checked-in raw results locally and compares them with their audited
-artifacts without making provider calls or writing new results.
+## Development from source
 
-## Maintainer release
-
-OpenMerit follows pi's [package format](https://github.com/earendil-works/pi/blob/main/packages/coding-agent/docs/packages.md):
-the npm manifest has the `pi-package` keyword, a `pi.extensions` entry, and the
-pi API as a peer dependency. Public npm packages with that keyword are eligible
-for pi's package gallery; a separate approval from the pi team is not part of
-the standard listing flow.
-
-Before publishing a release:
-
-```bash
-npm run check
-npm pack --dry-run
-npm publish --access public
+```sh
+npm ci
+npm run verify
+./node_modules/.bin/pi -e ./packages/pi/src/index.ts
 ```
 
-Then verify the actual public artifact with `npm view openmerit` and a clean
-`pi install npm:openmerit`. Gallery indexing may not be immediate. A featured
-or curated mention in pi-owned documentation is separate and would require the
-maintainers to accept a contribution or request.
+Credentials must never be committed. The repository intentionally contains no provider or Jev API key.

package/dist/core/src/index.d.ts

@@ -0,0 +1,90 @@
+import type { AnyIntentResult, AutomationPlan, AutomationSignal, CandidateAssessment, FrontierSnapshot, HarnessCapability, HarnessDescriptor, HarnessDispatchReceipt, HarnessWakeupRequest, TaskProfile, ImprovementCycleState, IntentLifecycleUpdate, IntentKind, OpenMeritIntent, OpenMeritProjectConfig, OpenMeritProjectState, ObservabilityCoverage, JsonValue, SupervisedGraduationPolicy } from "../../protocol/src/index.ts";
+import type { OpenMeritStore } from "./store.ts";
+export { OPENMERIT_DIRECTORY, ProjectStore, type AuditEventExporter, type OpenMeritStore, type ProjectStoreOptions, } from "./store.ts";
+export interface HarnessAdapter {
+    readonly descriptor: HarnessDescriptor;
+    dispatch(intent: OpenMeritIntent): Promise<HarnessDispatchReceipt>;
+    scheduleWakeup?(request: HarnessWakeupRequest): Promise<HarnessDispatchReceipt>;
+}
+export interface IntentFactoryOptions {
+    readonly taskProfileId?: string;
+    readonly authorizationPolicyId?: string;
+    readonly policy?: SupervisedGraduationPolicy;
+    readonly id?: string;
+    readonly requestedAt?: string;
+}
+export declare function createOpenMeritIntent(kind: IntentKind, options?: IntentFactoryOptions): OpenMeritIntent;
+export declare function missingHarnessCapabilities(descriptor: HarnessDescriptor, kind: IntentKind): readonly HarnessCapability[];
+export declare function validateHarnessDescriptor(descriptor: HarnessDescriptor): VerificationResult;
+export declare function dispatchToHarness(adapter: HarnessAdapter, intent: OpenMeritIntent): Promise<HarnessDispatchReceipt>;
+export declare function startedStage(kind: IntentKind): ImprovementCycleState["stage"];
+export declare function completedStage(kind: IntentKind, succeeded: boolean): ImprovementCycleState["stage"];
+export interface VerificationResult {
+    readonly valid: boolean;
+    readonly errors: readonly string[];
+}
+export interface ResultInterpretation extends VerificationResult {
+    readonly nextStage?: ImprovementCycleState["stage"];
+    readonly config?: OpenMeritProjectConfig;
+    readonly cyclePatch?: Partial<ImprovementCycleState>;
+}
+export declare function validateObservabilityCoverage(profile: TaskProfile, coverage: ObservabilityCoverage, requireReady: boolean): readonly string[];
+export declare function interpretIntentResult(intent: OpenMeritIntent, result: AnyIntentResult, cycle: ImprovementCycleState, config: OpenMeritProjectConfig): ResultInterpretation;
+export type OrchestrationDecision = {
+    readonly action: "issue_intent";
+    readonly intentKind: IntentKind;
+    readonly reason: string;
+} | {
+    readonly action: "await_user";
+    readonly reason: string;
+} | {
+    readonly action: "observe";
+    readonly reason: string;
+};
+export declare function nextImprovementAction(state: ImprovementCycleState): OrchestrationDecision;
+export declare function verifyFrontierSnapshot(profile: TaskProfile, assessments: readonly CandidateAssessment[], snapshot: FrontierSnapshot): VerificationResult;
+export interface IssuedIntent {
+    readonly intent: OpenMeritIntent;
+    readonly state: OpenMeritProjectState;
+    readonly receipt: HarnessDispatchReceipt;
+}
+export interface AcceptedIntentResult {
+    readonly intentKind: IntentKind;
+    readonly state: OpenMeritProjectState;
+    readonly interpretation: ResultInterpretation;
+    readonly automation: {
+        readonly plan: AutomationPlan;
+        readonly receipt?: HarnessDispatchReceipt;
+    };
+}
+export interface ObservationResult {
+    readonly state: OpenMeritProjectState;
+    readonly assessmentDue: boolean;
+}
+export interface CatalogChangeResult {
+    readonly state: OpenMeritProjectState;
+    readonly changed: boolean;
+    readonly reassessmentDue: boolean;
+}
+export interface AutomationSignalResult {
+    readonly state: OpenMeritProjectState;
+    readonly duplicate: boolean;
+    readonly decision: OrchestrationDecision;
+}
+export declare function buildAutomationPlan(state: OpenMeritProjectState, descriptor: HarnessDescriptor): AutomationPlan;
+/** Harness-neutral durable coordinator. Adapters transport jobs; this class owns lifecycle state. */
+export declare class OpenMeritCoordinator {
+    private readonly store;
+    private readonly adapter;
+    constructor(store: OpenMeritStore, adapter: HarnessAdapter);
+    issue(kind: IntentKind): Promise<IssuedIntent>;
+    reconcileAutomation(now?: string): Promise<{
+        readonly plan: AutomationPlan;
+        readonly receipt?: HarnessDispatchReceipt;
+    }>;
+    acceptResult(result: AnyIntentResult): Promise<AcceptedIntentResult>;
+    recordLifecycle(update: IntentLifecycleUpdate): Promise<void>;
+    recordTaskObservation(details?: Readonly<Record<string, JsonValue>>): Promise<ObservationResult>;
+    recordAutomationSignal(signal: AutomationSignal): Promise<AutomationSignalResult>;
+    recordModelCatalogFingerprint(fingerprint: string): Promise<CatalogChangeResult>;
+}