openplanr 1.25.3 → 2.0.0

package/dist/services/operate/advisors.js

@@ -1,1545 +0,0 @@
-import { randomUUID } from 'node:crypto';
-import { readFileSync } from 'node:fs';
-import { mkdir, readdir, readFile, rename, writeFile } from 'node:fs/promises';
-import path from 'node:path';
-import { pathToFileURL } from 'node:url';
-import { z } from 'zod';
-import { AIError } from '../../ai/errors.js';
-import { DEFAULT_MODELS } from '../../ai/types.js';
-import { OPENPLANR_VERSION } from '../../utils/package-version.js';
-import { getAIProvider, isAIConfigured } from '../ai-service.js';
-import { loadConfig } from '../config-service.js';
-import { isPlanrArtifactId } from './artifacts.js';
-import { canonicalDigest, canonicalize } from './canonical.js';
-import { buildOperatingBootstrapMap } from './context-research.js';
-import { createMissionToolset, MISSION_READ_ONLY_TOOLS, narrowMissionRootsToCeiling, operatingRuntimeEnforcesBoundedReadOnly, resolveOperatingDispatchIsolation, runMissionDispatchFanOut, } from './mission-dispatch.js';
-import { assertOperatingArtifact, loadOperatingProtocol, resolveOperatingPipelineRoot, } from './protocol.js';
-import { sanitizeGeneratedPlainText } from './redaction.js';
-import { OPERATE_PROTOCOL_VERSION, OPERATE_SCHEMA_VERSION, OperateError, } from './types.js';
-const proposalSchema = z
-    .object({
-    proposalKey: z.string().regex(/^[A-Za-z0-9._-]+$/),
-    type: z.enum(['finding', 'decision', 'data-gap', 'merge', 'sequence']),
-    title: z.string().min(1),
-    problem: z.string().min(1),
-    proposal: z.string().min(1),
-    impact: z.number().int().min(1).max(5),
-    confidence: z.number().int().min(1).max(5),
-    ease: z.number().int().min(1).max(5),
-    severity: z.enum(['low', 'medium', 'high', 'critical']),
-    evidenceRefs: z.array(z.string().regex(/^EVD-[A-Za-z0-9._-]+$/)).min(1),
-    dependsOnProposalKeys: z.array(z.string().regex(/^[A-Za-z0-9._-]+$/)).optional(),
-    conflictsWithProposalKeys: z.array(z.string().regex(/^[A-Za-z0-9._-]+$/)).optional(),
-    sequenceProposalKeys: z
-        .array(z.string().regex(/^[A-Za-z0-9._-]+$/))
-        .min(2)
-        .optional(),
-})
-    .strict();
-const advisorOutputSchema = z
-    .object({
-    outcome: z.enum(['proposals', 'quiet']),
-    proposals: z.array(proposalSchema).max(20),
-    gaps: z.array(z.string()),
-    conflicts: z.array(z.string()),
-})
-    .strict();
-// The mission proposal's citation shape, kept in lockstep with the anchor the
-// record-time resolver validates (`operating-citation@1.4.0`): a repository path
-// (relative, dot-prefixed roots allowed, no `..` traversal) / git revision /
-// planr artifact (any real artifact class), optionally scoped to a workspace
-// `componentId`, and bound to the cycle's frozen `pinnedRevision`. Each proposal
-// carries `citations` INSTEAD of the v1.2 `evidenceRefs`; the resolver snapshots
-// each citation after the lens returns and OpenPlanr never widens the set.
-const missionCitationSchema = z
-    .object({
-    citationKey: z
-        .string()
-        .regex(/^[A-Za-z0-9._-]+$/)
-        .max(128)
-        .optional(),
-    componentId: z
-        .string()
-        .regex(/^[a-z][a-z0-9-]{0,63}$/)
-        .optional(),
-    repositoryPath: z
-        .string()
-        .min(1)
-        .max(1024)
-        .regex(/^(?!\/)(?!.*(?:^|\/)\.\.(?:\/|$)).+$/)
-        .optional(),
-    lineRange: z
-        .object({ start: z.number().int().min(1), end: z.number().int().min(1) })
-        .strict()
-        .optional(),
-    gitRevision: z
-        .string()
-        .regex(/^[A-Fa-f0-9]{7,64}$/)
-        .optional(),
-    planrArtifactId: z
-        .string()
-        .refine(isPlanrArtifactId, 'must name a known planr artifact class')
-        .optional(),
-    pinnedRevision: z.string().regex(/^[A-Fa-f0-9]{7,64}$/),
-})
-    .strict();
-const missionProposalSchema = z
-    .object({
-    proposalKey: z.string().regex(/^[A-Za-z0-9._-]+$/),
-    type: z.enum(['finding', 'decision', 'data-gap', 'merge', 'sequence']),
-    title: z.string().min(1),
-    problem: z.string().min(1),
-    proposal: z.string().min(1),
-    impact: z.number().int().min(1).max(5),
-    confidence: z.number().int().min(1).max(5),
-    ease: z.number().int().min(1).max(5),
-    severity: z.enum(['low', 'medium', 'high', 'critical']),
-    citations: z.array(missionCitationSchema).min(1).max(50),
-    dependsOnProposalKeys: z.array(z.string().regex(/^[A-Za-z0-9._-]+$/)).optional(),
-    conflictsWithProposalKeys: z.array(z.string().regex(/^[A-Za-z0-9._-]+$/)).optional(),
-    sequenceProposalKeys: z
-        .array(z.string().regex(/^[A-Za-z0-9._-]+$/))
-        .min(2)
-        .optional(),
-})
-    .strict();
-const missionAdvisorOutputSchema = z
-    .object({
-    outcome: z.enum(['proposals', 'quiet']),
-    proposals: z.array(missionProposalSchema).max(20),
-    gaps: z.array(z.string()),
-    conflicts: z.array(z.string()),
-})
-    .strict();
-function v14CitationToMission(value, pinnedRevision) {
-    // A `componentId` scopes a repository/git locator to a sibling workspace
-    // component so it is resolved against that component's checkout, not the
-    // control repository. Preserve it through the conversion.
-    const componentId = typeof value.componentId === 'string' ? { componentId: value.componentId } : {};
-    if (value.kind === 'repository') {
-        return {
-            ...componentId,
-            repositoryPath: String(value.path),
-            lineRange: { start: Number(value.startLine), end: Number(value.endLine) },
-            pinnedRevision: String(value.revision),
-        };
-    }
-    if (value.kind === 'git') {
-        return {
-            ...componentId,
-            gitRevision: String(value.revision),
-            pinnedRevision: String(value.revision),
-        };
-    }
-    if (value.kind === 'planr') {
-        return {
-            planrArtifactId: String(value.artifactId),
-            pinnedRevision,
-        };
-    }
-    // External citations require a consented connected-research snapshot. They
-    // cannot be silently converted into a local repository citation.
-    return null;
-}
-/**
- * SINGLE source of truth for the action `routeKind` → committed proposal `type`
- * mapping. A Protocol v1.4 native ACTION carries a `routeKind`; the committed
- * proposal it becomes carries a `type` from the frozen proposal-type vocabulary.
- * Only a `decision` routeKind becomes a `decision` proposal; every other route
- * kind (`quick-task`, `spec`, `epic`, `agent-artifact`, `experiment`, `metric`)
- * becomes a `finding`. The registry-derived brief bounds
- * (`deriveRegistryProposalImage`), the native-response normalizer, and the
- * raw-response type projection all route through here so the three can never
- * disagree on what an action becomes — the exact divergence that let the Chair's
- * brief exclude `finding`/`decision` while every one of its actions mapped to
- * them, so the Chair could record ZERO actions and failed on any action at all.
- */
-export function routeKindToProposalType(routeKind) {
-    return routeKind === 'decision' ? 'decision' : 'finding';
-}
-/** The two registry facets that drive a role's proposal bounds. */
-const registryProposalBoundsSchema = z
-    .object({
-    allowedRouteKinds: z.array(z.string().min(1)).min(1),
-    budgets: z.object({ maxActions: z.number().int().positive() }),
-})
-    .passthrough();
-/**
- * The legacy protocol version OpenPlanr explicitly requests when it needs a role's
- * pack-style brief. The brief supplies ONLY the per-role bounds the record path
- * enforces (the proposal cap, the allowed proposal types, the output-byte ceiling)
- * and the capability tier — never the response contract, which a mandate discloses
- * from its own `responseSchema`. The version is stated rather than defaulted: the
- * pipeline refuses a Protocol v1.4 brief outright and no longer guesses a version,
- * because a silently defaulted one is exactly how a legacy contract reaches a
- * session that never asked for it.
- */
-const LEGACY_ADVISOR_BRIEF_PROTOCOL_VERSION = '1.2.0';
-/**
- * The IMAGE of a role's registry `allowedRouteKinds` under `routeKindToProposalType`
- * — exactly the proposal types an action of one of those route kinds becomes, i.e.
- * the set an advisor's actions can ACTUALLY produce — plus the registry's action
- * cap (`budgets.maxActions`). `merge`/`sequence`/`data-gap` have no routeKind
- * pre-image and are intentionally absent here (no action can reach them; making
- * them action-expressible is a protocol question, not this derivation's job).
- */
-function deriveRegistryProposalImage(role) {
-    const parsed = registryProposalBoundsSchema.safeParse(role);
-    if (!parsed.success) {
-        throw new OperateError('E_OPERATE_STATE_INVALID', `Operating role registry entry ${role.id} is missing the allowedRouteKinds / budgets.maxActions needed to derive its proposal bounds.`);
-    }
-    return {
-        allowedProposalTypes: [
-            ...new Set(parsed.data.allowedRouteKinds.map(routeKindToProposalType)),
-        ].sort(),
-        maximumProposals: parsed.data.budgets.maxActions,
-    };
-}
-/**
- * Build a role's advisor brief with its proposal BOUNDS reconciled against the
- * live role registry — the single source of truth — so the registry and the
- * enforced runtime contract can never disagree again. The pipeline's pack-style
- * brief is a frozen Protocol v1.2 compatibility projection whose
- * `output.allowedProposalTypes` predate the v1.4 action model: for the Chair it is
- * `['merge','sequence']`, DISJOINT from the set every one of the Chair's actions
- * maps to (`finding`/`decision`). So the Chair could commit a quiet result but was
- * rejected the instant it proposed a single bounded route — the exact defect this
- * closes.
- *
- * Reconciliation:
- *  - `maximumProposals` is taken from the registry's `budgets.maxActions` (today
- *    already identical to the pipeline value for every role, but now set
- *    explicitly so a future registry cap edit cannot silently disagree with the
- *    runtime).
- *  - `allowedProposalTypes` is the UNION of the registry image and the frozen
- *    brief's existing types. The union ADDS the registry-reachable action types
- *    (`finding`/`decision`) — which is what unblocks the Chair — while LEAVING the
- *    frozen structural consolidation vocabulary (`merge`/`sequence` for the Chair;
- *    `data-gap` for the lenses) untouched: those have no action pre-image, so they
- *    are neither invented as action-reachable nor removed as capabilities. For
- *    every non-Chair role the image is already a subset of the frozen set, so the
- *    reconciliation is a byte-identical no-op there.
- *
- * The brief digest is recomputed over the reconciled content so the artifact stays
- * internally consistent; `canonicalize` here is the same RFC 8785 JCS the pipeline
- * signs briefs with, so a no-op reconciliation reproduces the pipeline digest
- * exactly.
- */
-export function createRegistryReconciledAdvisorBrief(protocol, roleId) {
-    const brief = protocol.createOperatingAdvisorBrief(roleId, {
-        protocolVersion: LEGACY_ADVISOR_BRIEF_PROTOCOL_VERSION,
-    });
-    const role = protocol.listOperatingRoles().find((entry) => entry.id === roleId);
-    if (!role) {
-        throw new OperateError('E_OPERATE_STATE_INVALID', `Operating role ${roleId} is absent from the registry; its proposal bounds cannot be reconciled.`);
-    }
-    const image = deriveRegistryProposalImage(role);
-    const allowedProposalTypes = [
-        ...new Set([
-            ...brief.output.allowedProposalTypes,
-            ...image.allowedProposalTypes,
-        ]),
-    ].sort();
-    const { briefDigest: _priorDigest, ...unsigned } = {
-        ...brief,
-        output: {
-            ...brief.output,
-            allowedProposalTypes,
-            maximumProposals: image.maximumProposals,
-        },
-    };
-    return { ...unsigned, briefDigest: canonicalDigest(unsigned) };
-}
-export function normalizeAgentNativeResponse(response, pinnedRevision) {
-    return {
-        outcome: response.actions.length > 0 ? 'proposals' : 'quiet',
-        proposals: response.actions.map((action) => ({
-            proposalKey: action.actionKey,
-            type: routeKindToProposalType(action.routeKind),
-            title: action.title,
-            problem: action.summary,
-            proposal: action.summary,
-            impact: action.impact ?? 3,
-            confidence: action.confidence,
-            ease: action.ease ?? 3,
-            severity: action.critical
-                ? 'critical'
-                : (action.impact ?? 3) >= 4
-                    ? 'high'
-                    : (action.impact ?? 3) >= 3
-                        ? 'medium'
-                        : 'low',
-            citations: action.citations
-                .map((citation) => v14CitationToMission(citation, pinnedRevision))
-                .filter((citation) => citation !== null),
-        })),
-        gaps: response.gaps.map((gap) => `${gap.question} Impact: ${gap.impact}`),
-        conflicts: response.conflicts.map((conflict) => conflict.commitmentRef
-            ? `${conflict.summary} — conflicts with the published commitment "${conflict.commitmentRef.statement}" (${conflict.commitmentRef.path})`
-            : conflict.summary),
-    };
-}
-export function advisorResponseContractDetails(brief) {
-    const examples = brief.output.jsonSchema?.examples;
-    return {
-        expectedSchema: 'operating-advisor-response@1.2.0',
-        example: Array.isArray(examples)
-            ? examples
-            : [{ outcome: 'quiet', proposals: [], gaps: [], conflicts: [] }],
-    };
-}
-/**
- * SINGLE source of truth for the per-role proposal bounds — the proposal cap and
- * the allowed proposal types — over an already-projected `{type}` list. Both the
- * committed-output guard (`collectBriefContractViolations`) and the raw-response
- * batch validator (`collectAdvisorResponseContractIssues`) route through here so
- * the record path and the `harness validate` dry-run can never disagree on them.
- */
-function collectProposalBoundViolations(brief, proposals) {
-    const issues = [];
-    if (proposals.length > brief.output.maximumProposals) {
-        issues.push({
-            path: 'proposals',
-            rule: 'maximumProposals',
-            detail: `Advisor ${brief.role.id} returned ${proposals.length} proposals; its canonical limit is ${brief.output.maximumProposals}.`,
-        });
-    }
-    const allowed = new Set(brief.output.allowedProposalTypes);
-    const disallowed = [
-        ...new Set(proposals
-            .map((proposal) => proposal.type)
-            .filter((type) => typeof type === 'string' && !allowed.has(type))),
-    ].sort();
-    if (disallowed.length > 0) {
-        issues.push({
-            path: 'proposals',
-            rule: 'allowedProposalTypes',
-            detail: `Advisor ${brief.role.id} returned proposal types outside its canonical brief: ${disallowed.join(', ')}.`,
-        });
-    }
-    return issues;
-}
-/**
- * Collect EVERY brief-level violation of a NORMALIZED advisor output — the
- * proposal cap, disallowed proposal types, outcome/proposal consistency, and the
- * output-byte ceiling — as `{path, rule, detail}` issues instead of throwing on
- * the first. `assertAdvisorOutputMatchesBrief` wraps this to keep its throwing
- * contract at the committed-output guards.
- */
-export function collectBriefContractViolations(brief, output) {
-    const issues = collectProposalBoundViolations(brief, output.proposals);
-    if (output.outcome === 'quiet' && output.proposals.length > 0) {
-        issues.push({
-            path: 'outcome',
-            rule: 'outcome-consistency',
-            detail: `Advisor ${brief.role.id} declared a quiet result with proposals.`,
-        });
-    }
-    if (output.outcome === 'proposals' && output.proposals.length === 0) {
-        issues.push({
-            path: 'outcome',
-            rule: 'outcome-consistency',
-            detail: `Advisor ${brief.role.id} declared proposals without a proposal.`,
-        });
-    }
-    const outputBytes = Buffer.byteLength(canonicalize(output), 'utf8');
-    if (outputBytes > brief.output.maximumOutputBytes) {
-        issues.push({
-            path: 'output',
-            rule: 'maximumOutputBytes',
-            detail: `Advisor ${brief.role.id} exceeded its canonical output budget.`,
-        });
-    }
-    return issues;
-}
-export function assertAdvisorOutputMatchesBrief(brief, output) {
-    const violations = collectBriefContractViolations(brief, output);
-    if (violations.length > 0) {
-        throw new OperateError('E_OPERATE_ADVISOR_FAILED', violations.length === 1
-            ? violations[0].detail
-            : `Advisor ${brief.role.id} output violates its canonical brief in ${violations.length} ways: ${violations
-                .map((violation) => violation.detail)
-                .join(' ')}`, { issues: violations });
-    }
-}
-/**
- * Project the proposal `{type}` list from a RAW submitted advisor response so the
- * per-role cap and allowed-type checks can run even on a schema-invalid payload —
- * a schema failure must never short-circuit the cap/type categories. v1.4 native
- * responses carry `actions` whose `routeKind` maps to a proposal type exactly as
- * `normalizeAgentNativeResponse` does (only `decision` is a decision proposal);
- * v1.3 responses carry `proposals` with an explicit `type`.
- */
-function projectSubmittedProposalTypes(response, responseProtocol) {
-    const record = response && typeof response === 'object' && !Array.isArray(response)
-        ? response
-        : {};
-    if (responseProtocol === '1.4.0') {
-        const actions = Array.isArray(record.actions) ? record.actions : [];
-        return actions.map((action) => {
-            const routeKind = action && typeof action === 'object' && !Array.isArray(action)
-                ? action.routeKind
-                : undefined;
-            return { type: routeKindToProposalType(typeof routeKind === 'string' ? routeKind : '') };
-        });
-    }
-    const proposals = Array.isArray(record.proposals) ? record.proposals : [];
-    return proposals.map((proposal) => {
-        const type = proposal && typeof proposal === 'object' && !Array.isArray(proposal)
-            ? proposal.type
-            : undefined;
-        return { type: typeof type === 'string' ? type : undefined };
-    });
-}
-/**
- * US-T1 batch validation: collect ALL advisor-response contract violations in one
- * pass — the response-schema issues (a bad enum, a wrong-typed `gaps[].impact`, a
- * malformed citation shape, …), the per-role proposal cap, and disallowed proposal
- * types — as a single `{path, rule, detail}` list. Shared verbatim by the `record`
- * path and the `harness validate` dry-run so the two can never diverge; the caller
- * decides whether a non-empty list is a rejection (record) or a dry-run report
- * (validate). Reads the RAW submitted response so a schema-invalid payload still
- * discloses its cap/type violations in the same response instead of forcing a
- * category-at-a-time resubmission cycle.
- */
-export async function collectAdvisorResponseContractIssues(input) {
-    const protocol = await loadOperatingProtocol();
-    const responseProtocol = input.protocolVersion === '1.4.0' ? '1.4.0' : '1.3.0';
-    const issues = [];
-    // Category 1 — response schema (bad enum, wrong-typed gap impact, malformed
-    // citation shape, missing required fields). The protocol validator already
-    // returns EVERY schema issue at once; we only stop it from being the sole
-    // category by continuing to the brief-level bounds below.
-    for (const issue of protocol.validateProtocolArtifact('operating-advisor-response', input.response, { protocolVersion: responseProtocol })) {
-        // Preserve the pipeline's specific failing rule (`enum`, `oneOf`, `type`, …)
-        // rather than flattening every schema failure to one label, so a runtime sees
-        // exactly which constraint each path violated.
-        issues.push({ path: issue.path, rule: issue.rule, detail: issue.detail });
-    }
-    // Categories 2 & 3 — per-role proposal cap + disallowed proposal types, read
-    // from the raw response so they surface alongside any schema issues.
-    issues.push(...collectProposalBoundViolations(input.brief, projectSubmittedProposalTypes(input.response, responseProtocol)));
-    return issues;
-}
-const CHARTER_FIELDS_BY_ROLE = {
-    'strategy-finance': [
-        'purpose',
-        'stage',
-        'businessModel',
-        'goals',
-        'constraints',
-        'successMetrics',
-    ],
-    'technology-risk': ['purpose', 'stage', 'goals', 'constraints', 'guardrails', 'knownUnknowns'],
-    'product-activation': [
-        'purpose',
-        'stage',
-        'idealCustomer',
-        'goals',
-        'successMetrics',
-        'knownUnknowns',
-    ],
-    'growth-market': [
-        'purpose',
-        'stage',
-        'businessModel',
-        'idealCustomer',
-        'goals',
-        'successMetrics',
-    ],
-    'operations-customer': [
-        'purpose',
-        'stage',
-        'idealCustomer',
-        'goals',
-        'constraints',
-        'guardrails',
-        'knownUnknowns',
-    ],
-    chair: [
-        'purpose',
-        'stage',
-        'businessModel',
-        'idealCustomer',
-        'goals',
-        'constraints',
-        'successMetrics',
-        'guardrails',
-        'knownUnknowns',
-    ],
-};
-function boundedContextText(value, maximum = 1_024) {
-    const text = typeof value === 'string' ? value : '';
-    return sanitizeGeneratedPlainText(text).replace(/\s+/g, ' ').trim().slice(0, maximum);
-}
-function charterSection(markdown, heading) {
-    const normalized = markdown.replace(/\r\n?/g, '\n');
-    const marker = `## ${heading}`;
-    const start = normalized.indexOf(marker);
-    if (start < 0)
-        return [];
-    const body = normalized.slice(start + marker.length);
-    const end = body.search(/\n##\s+/);
-    return (end >= 0 ? body.slice(0, end) : body)
-        .split('\n')
-        .map((line) => line.match(/^\s*-\s+(.+?)\s*$/)?.[1] ?? '')
-        .map((line) => boundedContextText(line))
-        .filter((line) => line && !line.startsWith('[unknown') && !line.startsWith('[none recorded'));
-}
-function parseOperatingCharter(markdown) {
-    const product = charterSection(markdown, 'Product context');
-    const productValues = new Map(product.flatMap((line) => {
-        const match = line.match(/^([^:]+):\s*(.*)$/);
-        return match ? [[match[1].trim().toLowerCase(), match[2].trim()]] : [];
-    }));
-    return {
-        purpose: productValues.get('purpose') ?? '',
-        stage: productValues.get('stage') ?? '',
-        businessModel: productValues.get('business model') ?? '',
-        idealCustomer: productValues.get('ideal customer') ?? '',
-        goals: charterSection(markdown, 'Current goals').sort(),
-        constraints: charterSection(markdown, 'Constraints').sort(),
-        successMetrics: charterSection(markdown, 'Success metrics').sort(),
-        guardrails: charterSection(markdown, 'Guardrails').sort(),
-        knownUnknowns: charterSection(markdown, 'Known unknowns').sort(),
-    };
-}
-function recordEvidenceRefs(record) {
-    return Array.isArray(record.evidenceRefs)
-        ? [
-            ...new Set(record.evidenceRefs.filter((reference) => typeof reference === 'string')),
-        ].sort()
-        : [];
-}
-function openItem(record, summaryKeys) {
-    const summary = summaryKeys.map((key) => boundedContextText(record[key])).find(Boolean) ?? 'Review required.';
-    return {
-        id: record.id,
-        status: record.status,
-        summary,
-        owner: boundedContextText(record.owner) || null,
-        evidenceRefs: recordEvidenceRefs(record),
-        ...(Array.isArray(record.affectedRoles)
-            ? {
-                affectedRoles: [
-                    ...new Set(record.affectedRoles.filter((role) => typeof role === 'string')),
-                ].sort(),
-            }
-            : {}),
-    };
-}
-/** Build the immutable non-advisor context shared by all independent lenses. */
-export async function buildAdvisorOperatingContext(input) {
-    const charter = parseOperatingCharter(await readFile(input.charterPath, 'utf8'));
-    const prior = [...input.state.cycles]
-        .filter((cycle) => cycle.id !== input.cycleId)
-        .sort((left, right) => String(left.id).localeCompare(String(right.id)))
-        .at(-1);
-    const priorCycle = prior
-        ? {
-            id: prior.id,
-            state: prior.state,
-            health: typeof prior.health === 'string' ? prior.health : null,
-            findings: input.state.findings.filter((finding) => finding.cycleId === prior.id).length,
-            decisions: input.state.decisions.filter((decision) => decision.cycleId === prior.id).length,
-            gaps: input.state.dataGaps.filter((gap) => gap.cycleId === prior.id).length,
-            pendingOutcomes: input.state.outcomes.filter((outcome) => outcome.sourceCycle === prior.id && outcome.status === 'pending').length,
-        }
-        : null;
-    const openDecisions = input.state.decisions
-        .filter((decision) => ['open', 'default-due'].includes(decision.status))
-        .map((decision) => openItem(decision, ['question', 'recommendation']))
-        .sort((left, right) => left.id.localeCompare(right.id));
-    const openGaps = input.state.dataGaps
-        .filter((gap) => ['open', 'answered'].includes(gap.status))
-        .map((gap) => openItem(gap, ['question', 'reason']))
-        .sort((left, right) => left.id.localeCompare(right.id));
-    const pendingOutcomes = input.state.outcomes
-        .filter((outcome) => ['pending', 'observing', 'inconclusive'].includes(outcome.status))
-        .map((outcome) => openItem(outcome, ['metric', 'queryIdentity']))
-        .sort((left, right) => left.id.localeCompare(right.id));
-    const unsigned = {
-        charter,
-        priorCycle,
-        openDecisions,
-        openGaps,
-        pendingOutcomes,
-    };
-    return { ...unsigned, snapshotDigest: canonicalDigest(unsigned) };
-}
-function roleContext(context, roleId, permittedEvidenceRefs) {
-    const charter = Object.fromEntries(CHARTER_FIELDS_BY_ROLE[roleId].map((field) => [field, structuredClone(context.charter[field])]));
-    const visible = (item) => {
-        if (roleId === 'chair')
-            return true;
-        if (item.affectedRoles && item.affectedRoles.length > 0) {
-            return item.affectedRoles.includes(roleId);
-        }
-        return (item.evidenceRefs.length === 0 ||
-            item.evidenceRefs.some((reference) => permittedEvidenceRefs.has(reference)));
-    };
-    const filtered = {
-        charter,
-        priorCycle: context.priorCycle,
-        openDecisions: context.openDecisions.filter(visible),
-        openGaps: context.openGaps.filter(visible),
-        pendingOutcomes: context.pendingOutcomes.filter(visible),
-    };
-    return { ...filtered, snapshotDigest: canonicalDigest(filtered) };
-}
-/**
- * Bounded fan-out concurrency for advisor dispatch (FR12): a fixed pool caps how
- * many lenses research at once without limiting any single lens's depth.
- */
-export const DEFAULT_OPERATING_DISPATCH_CONCURRENCY = 3;
-/**
- * Default per-role research time budget (FR12). Generous by design: it is a signal
- * to STOP researching and synthesize, never a cut-off, so a role that reaches it
- * still returns its full analysis. Raising or removing it never truncates output.
- */
-export const DEFAULT_OPERATING_ROLE_RESEARCH_BUDGET_MS = 8 * 60_000;
-/**
- * Compose a role's FR12 research guidance from the shared bootstrap map and its
- * registry-derived investigation mandate. The stop criteria are explicit prose so
- * the lens knows WHEN to stop reading and synthesize; the budget line always states
- * that reaching the budget means concluding from what was gathered, never dropping
- * or truncating a conclusion to fit it.
- */
-function buildMandateResearchGuidance(mandate, bootstrapMap, researchBudgetMs) {
-    const budgetMs = researchBudgetMs && researchBudgetMs > 0 ? researchBudgetMs : null;
-    const deduplicationHints = [
-        'A shared bootstrap map already indexes this project’s planning families and recent git history; use its citations as entry points instead of re-walking the workspace to rediscover them.',
-        ...(bootstrapMap?.searchHints ?? []),
-    ];
-    const stopResearchingAndSynthesize = [
-        'Stop once every material claim and action is grounded in a resolvable citation and further reading no longer changes your ranked conclusions.',
-        'Stop once you have reconciled the shared bootstrap map’s cited planning and git indexes for your focus areas against a direct reading of the relevant sources.',
-        budgetMs
-            ? `When your per-role research budget of ~${Math.round(budgetMs / 1000)}s elapses, synthesize from what you have gathered — the budget is a signal to conclude, never a cut-off. Return your full analysis and every citation-qualified action, and never drop or truncate a conclusion to fit it.`
-            : 'When you reach your research time budget, synthesize from what you have gathered rather than continuing to read — never drop or truncate a conclusion to fit it.',
-    ];
-    return {
-        bootstrapMap,
-        bootstrapMapDigest: bootstrapMap?.mapDigest ?? null,
-        focusAreas: [...(mandate.investigationMandate?.examine ?? [])],
-        deduplicationHints,
-        perRoleResearchBudgetMs: budgetMs,
-        stopResearchingAndSynthesize,
-    };
-}
-let cachedMandateApi = null;
-async function loadOperatingMandateApi() {
-    cachedMandateApi ??= (async () => {
-        const root = resolveOperatingPipelineRoot({ requireMission: true });
-        if (!root) {
-            throw new OperateError('E_OPERATE_MISSION_UNAVAILABLE', 'Mandate dispatch requires the pipeline package with Protocol v1.3 (operating mandate).');
-        }
-        const loaded = (await import(pathToFileURL(path.join(root, 'lib', 'operate', 'mandate.mjs')).href));
-        if (typeof loaded.createOperatingMandate !== 'function') {
-            throw new OperateError('E_PIPELINE_VERSION_INCOMPATIBLE', 'Installed planr-pipeline does not export the Protocol v1.3 operating mandate builder.');
-        }
-        return loaded;
-    })();
-    return cachedMandateApi;
-}
-/**
- * Build the Protocol v1.3 operating mandate for a role (FR1). The mandate's
- * boundaries are declared directly from the caller's granted workspace roots —
- * never an evidence-index-derived subset — so a gitignored `.planr/` tree is
- * fully readable when the caller declares it. The registry supplies the lens
- * question, investigation mandate, and sensitivity ceiling; this function only
- * threads the declared boundaries through.
- */
-export async function buildOperatingMandate(input) {
-    const api = await loadOperatingMandateApi();
-    const mandate = api.createOperatingMandate(input.roleId, {
-        roots: [...input.roots],
-        forbiddenPaths: [...(input.forbiddenPaths ?? [])],
-        runtime: input.runtime,
-        protocolVersion: input.protocolVersion,
-    });
-    // Leave the harness/compatibility path byte-identical: research guidance is
-    // attached only when a bootstrap map or a per-role budget is threaded in.
-    if (input.bootstrapMap === undefined && input.researchBudgetMs === undefined) {
-        return mandate;
-    }
-    return {
-        ...mandate,
-        researchGuidance: buildMandateResearchGuidance(mandate, input.bootstrapMap ?? null, input.researchBudgetMs ?? null),
-    };
-}
-const OPERATING_RESPONSE_SCHEMA_FILE = 'operating-advisor-response.schema.json';
-function schemaRecord(value) {
-    return value && typeof value === 'object' && !Array.isArray(value)
-        ? value
-        : null;
-}
-/**
- * Read one packaged protocol schema document from the resolved pipeline install.
- * Synchronous because the disclosure is attached on the synchronous prepare path;
- * a missing or unreadable schema is a typed failure, never a silent fallback to a
- * schema the runtime is not validated against.
- */
-function readPipelineProtocolSchema(protocolVersion, file) {
-    const root = resolveOperatingPipelineRoot({ requireMission: true });
-    if (!root) {
-        throw new OperateError('E_OPERATE_MISSION_UNAVAILABLE', 'Disclosing a mandate response contract requires the pipeline package with the Protocol v1.4 schemas.');
-    }
-    const location = path.join(root, 'schemas', `v${protocolVersion}`, file);
-    let raw;
-    try {
-        raw = readFileSync(location, 'utf8');
-    }
-    catch (error) {
-        throw new OperateError('E_PIPELINE_VERSION_INCOMPATIBLE', `Installed planr-pipeline does not ship ${file} for Protocol ${protocolVersion} at ${location}: ${error instanceof Error ? error.message : String(error)}`);
-    }
-    const parsed = schemaRecord(JSON.parse(raw));
-    if (!parsed) {
-        throw new OperateError('E_PIPELINE_VERSION_INCOMPATIBLE', `${location} is not a JSON Schema object, so the mandate cannot disclose it.`);
-    }
-    return parsed;
-}
-/** Collect every sibling-file `$ref` (`#/$defs/...` pointers are already local). */
-function collectRelativeSchemaRefs(node, into) {
-    if (Array.isArray(node)) {
-        for (const item of node)
-            collectRelativeSchemaRefs(item, into);
-        return;
-    }
-    const record = schemaRecord(node);
-    if (!record)
-        return;
-    for (const [key, value] of Object.entries(record)) {
-        if (key === '$ref' && typeof value === 'string') {
-            if (value.endsWith('.schema.json') && !value.includes('://'))
-                into.add(value);
-            continue;
-        }
-        collectRelativeSchemaRefs(value, into);
-    }
-}
-const disclosedResponseSchemas = new Map();
-/**
- * Resolve the advisor-response schema for one protocol version as a SELF-CONTAINED
- * document. The packaged v1.3/v1.4 response schemas reference the sibling citation
- * schema by file name, which a runtime holding only the mandate cannot dereference,
- * so each referenced resource is embedded under `$defs` verbatim. JSON Schema
- * 2020-12 bundling: an embedded resource keeps its own `$id`, so the existing
- * sibling `$ref` resolves against the embedded copy and that copy's internal
- * `#/$defs/...` pointers still resolve inside it. Nothing is rewritten, so the
- * disclosed schema stays byte-faithful to the enforced one.
- */
-function resolveDisclosedResponseSchema(protocolVersion) {
-    const cached = disclosedResponseSchemas.get(protocolVersion);
-    if (cached)
-        return structuredClone(cached);
-    const schema = readPipelineProtocolSchema(protocolVersion, OPERATING_RESPONSE_SCHEMA_FILE);
-    const identifier = typeof schema.$id === 'string' ? schema.$id : '';
-    if (!identifier.endsWith(`/v${protocolVersion}/${OPERATING_RESPONSE_SCHEMA_FILE}`)) {
-        throw new OperateError('E_PIPELINE_VERSION_INCOMPATIBLE', `The advisor-response schema packaged for Protocol ${protocolVersion} declares $id "${identifier}", which does not name that version; disclosing it would hand the runtime a contract it is not validated against.`);
-    }
-    const references = new Set();
-    collectRelativeSchemaRefs(schema, references);
-    const embedded = { ...(schemaRecord(schema.$defs) ?? {}) };
-    for (const reference of [...references].sort()) {
-        embedded[reference.replace(/\.schema\.json$/, '')] = readPipelineProtocolSchema(protocolVersion, reference);
-    }
-    const bundled = references.size > 0 ? { ...schema, $defs: embedded } : schema;
-    disclosedResponseSchemas.set(protocolVersion, bundled);
-    return structuredClone(bundled);
-}
-/** The `actions[].routeKind` vocabulary a Protocol v1.4 response may author. */
-function disclosedRouteKinds(jsonSchema) {
-    const routeKind = schemaRecord(schemaRecord(schemaRecord(schemaRecord(schemaRecord(jsonSchema.properties)?.actions)?.items)?.properties)?.routeKind);
-    const values = Array.isArray(routeKind?.enum) ? routeKind.enum : [];
-    const kinds = values.filter((value) => typeof value === 'string');
-    if (kinds.length === 0) {
-        throw new OperateError('E_PIPELINE_VERSION_INCOMPATIBLE', 'The Protocol v1.4 advisor-response schema declares no actions[].routeKind vocabulary, so the mandate cannot disclose the proposal types its actions can produce.');
-    }
-    return kinds;
-}
-/**
- * The proposal types the disclosed response can ACTUALLY produce. A Protocol v1.4
- * response carries no proposal `type` at all — it carries actions whose
- * `routeKind` the record path projects through `routeKindToProposalType` — so
- * disclosing the frozen v1.2 vocabulary (`data-gap`, or the Chair's
- * `merge`/`sequence`) would name types no v1.4 action can express. The disclosed
- * set is therefore the image of the disclosed route-kind vocabulary, and it must
- * stay within the set `record` enforces from the registry-reconciled brief; a set
- * the enforcer would reject is a defect, not a disclosure.
- */
-function disclosedProposalTypes(mandate, brief, jsonSchema) {
-    if (mandate.protocolVersion !== '1.4.0')
-        return [...brief.output.allowedProposalTypes];
-    const image = [...new Set(disclosedRouteKinds(jsonSchema).map(routeKindToProposalType))].sort();
-    const enforced = new Set(brief.output.allowedProposalTypes);
-    const rejected = image.filter((type) => !enforced.has(type));
-    if (rejected.length > 0) {
-        throw new OperateError('E_OPERATE_STATE_INVALID', `Advisor ${mandate.roleId} would disclose proposal types its own enforced brief rejects: ${rejected.join(', ')}.`);
-    }
-    return image;
-}
-/**
- * US-T1: ship the disclosed response contract INSIDE the dispatched mandate — the
- * dereferenceable response JSON Schema plus the per-role bounds `record` enforces
- * (the proposal cap, the reachable proposal types, the output-byte ceiling, and
- * required behaviors) — so a runtime can satisfy the contract instead of guessing
- * at a `responseSchema` name it cannot resolve.
- *
- * DEFECT this closes: the schema was previously taken from the role's pack-style
- * brief. That brief is a compatibility-only projection the pipeline builds for
- * Protocol v1.2/v1.3 sessions and refuses to build for v1.4 at all, so it can
- * never express the v1.4 contract — yet a v1.4 mandate enforcing
- * `operating-advisor-response@1.4.0` disclosed that legacy contract, and an advisor
- * that followed the disclosure literally was rejected by the enforced schema
- * (v1.2 `proposals` where v1.4 requires `analysisMarkdown`/`claims`/`actions`, and
- * a `proposals|quiet` outcome where v1.4 admits `actions|quiet|partial`). The
- * schema is therefore resolved from the mandate's OWN `responseSchema`, which is
- * also the value disclosed as `output.schema` — so the disclosed identity and the
- * enforced one agree by construction, at the type level as well as at runtime.
- *
- * Layered on AFTER the pipeline signs `mandateDigest` (the returned object keeps
- * the pipeline's signed digest untouched), exactly like `researchGuidance`, so no
- * digest recomputation is required and writer/verifier stay in agreement over the
- * immutable, signed mandate the digest pins.
- */
-export function attachMandateResponseContract(mandate, brief) {
-    const version = mandate.responseSchema.split('@')[1] ?? '';
-    const jsonSchema = resolveDisclosedResponseSchema(version);
-    return {
-        ...mandate,
-        output: {
-            schema: mandate.responseSchema,
-            jsonSchema,
-            allowedProposalTypes: disclosedProposalTypes(mandate, brief, jsonSchema),
-            maximumProposals: brief.output.maximumProposals,
-            // The ENFORCED ceiling, not the larger v1.4 registry budget: `record`
-            // measures the canonicalized committed output against this brief value, so
-            // disclosing a wider ceiling would recreate the divergence above.
-            maximumOutputBytes: brief.output.maximumOutputBytes,
-            requiredBehavior: [...brief.output.requiredBehavior],
-        },
-    };
-}
-// A mandate's `boundaries.roots` items must satisfy this pattern (leading dot
-// permitted, no `..` traversal), so `.planr` is a valid declared root.
-const INLINE_MANDATE_ROOT_PATTERN = /^(?!.*\.\.)[A-Za-z0-9._][A-Za-z0-9._/-]*$/;
-/**
- * FR1/FR2: derive a mandate's declared read roots directly from the project
- * working tree — every top-level directory the agent may traverse — for the
- * inline dispatch path. `.planr/` is ALWAYS declared, whether or not it exists
- * yet and REGARDLESS of git tracking, because the mission tool surface walks the
- * filesystem directly rather than a tracked-file inventory, so a gitignored `.planr/` tree
- * is fully readable (finding 2's tracked-file gap cannot reproduce here).
- * `.git`/`node_modules` are never declared, and every root is filtered to the
- * mandate's schema pattern. Mirrors the native adapter path's derivation so both
- * dispatch paths declare identical boundaries.
- */
-export async function deriveOperatingMandateRoots(projectRoot) {
-    const entries = await readdir(projectRoot, { withFileTypes: true }).catch(() => []);
-    const roots = new Set(['.planr']);
-    for (const entry of entries) {
-        if (!entry.isDirectory())
-            continue;
-        if (entry.name === '.git' || entry.name === 'node_modules')
-            continue;
-        if (INLINE_MANDATE_ROOT_PATTERN.test(entry.name))
-            roots.add(entry.name);
-    }
-    return [...roots].sort();
-}
-/**
- * Convert a compact native-harness response into the canonical, digest-bound
- * Protocol result. Runtime adapters should not manufacture protocol metadata,
- * producer fields, or JCS digests themselves.
- */
-function sanitizeMissionOutput(output) {
-    return {
-        outcome: output.outcome,
-        proposals: output.proposals
-            .map((proposal) => ({
-            ...proposal,
-            title: sanitizeGeneratedPlainText(proposal.title).replace(/\s+/g, ' ').trim(),
-            problem: sanitizeGeneratedPlainText(proposal.problem).replace(/\s+/g, ' ').trim(),
-            proposal: sanitizeGeneratedPlainText(proposal.proposal).replace(/\s+/g, ' ').trim(),
-            citations: [...proposal.citations],
-            ...(proposal.dependsOnProposalKeys
-                ? {
-                    dependsOnProposalKeys: [...new Set(proposal.dependsOnProposalKeys)].sort(),
-                }
-                : {}),
-            ...(proposal.conflictsWithProposalKeys
-                ? {
-                    conflictsWithProposalKeys: [...new Set(proposal.conflictsWithProposalKeys)].sort(),
-                }
-                : {}),
-            ...(proposal.sequenceProposalKeys
-                ? { sequenceProposalKeys: [...proposal.sequenceProposalKeys] }
-                : {}),
-        }))
-            .sort((left, right) => left.proposalKey.localeCompare(right.proposalKey) ||
-            left.type.localeCompare(right.type) ||
-            left.problem.localeCompare(right.problem)),
-        gaps: [...new Set(output.gaps.map(sanitizeGeneratedPlainText))].sort(),
-        conflicts: [...new Set(output.conflicts.map(sanitizeGeneratedPlainText))].sort(),
-    };
-}
-/**
- * The mandate's response contract is `operating-advisor-response@1.3.0`, whose
- * output facet mirrors the v1.2 brief's contract (allowed proposal types,
- * maxima), so a v1.3 response is validated against exactly the same invariants a
- * pack response is — reusing the pipeline's registry-derived brief as the single
- * source of truth.
- *
- * FR2: when the response's citations resolve to ZERO evidence IDs across every
- * proposal, the role commits `not_evaluated` — a schema-legal `quiet` result
- * (the frozen v1.2 role-result schema admits no `not_evaluated` outcome; the
- * not_evaluated state lives in the governed gap + the integrity surface) — and
- * the governed empty-grounding gap the citation gate opened is returned, with
- * `notEvaluated: true`, rather than a `quiet` that pretends the lens evaluated.
- */
-export async function createNativeMissionOperatingRoleResult(input) {
-    const protocol = await loadOperatingProtocol();
-    const responseProtocol = input.mandate.protocolVersion === '1.4.0' ? '1.4.0' : '1.3.0';
-    // Compact responses carry no protocol envelope, so select the schema from the
-    // immutable mandate instead of allowing additive resolution to guess.
-    const contractIssues = protocol.validateProtocolArtifact('operating-advisor-response', input.response, { protocolVersion: responseProtocol });
-    if (contractIssues.length > 0) {
-        throw new OperateError('E_OPERATE_ADVISOR_FAILED', `Native ${input.mandate.roleId} response does not match operating-advisor-response@${responseProtocol}.`, { issues: contractIssues.slice(0, 8) });
-    }
-    let output;
-    if (responseProtocol === '1.4.0') {
-        const response = input.response;
-        output = sanitizeMissionOutput(normalizeAgentNativeResponse(response, input.pinnedRevision ?? '0000000'));
-    }
-    else {
-        const parsed = missionAdvisorOutputSchema.safeParse(input.response);
-        if (!parsed.success) {
-            throw new OperateError('E_OPERATE_INTERNAL', 'Protocol and OpenPlanr disagree on the v1.3 mission advisor response contract.', {
-                issues: parsed.error.issues.slice(0, 8).map((issue) => ({
-                    path: issue.path.join('.'),
-                    code: issue.code,
-                })),
-            });
-        }
-        output = sanitizeMissionOutput(parsed.data);
-    }
-    const brief = createRegistryReconciledAdvisorBrief(protocol, input.mandate.roleId);
-    assertAdvisorOutputMatchesBrief(brief, output);
-    const capability = brief.role.capabilityTier;
-    // The intermediate result: proposals carry their v1.3 citations and an empty
-    // evidenceRefs set. It is deliberately NOT yet a v1.2-valid committed
-    // operating-role-result — the citation gate mints the evidenceRefs that make
-    // it one. `inputDigest` is the mandate's digest, so the record path's
-    // input-digest binding (prepare stored the same mandate digest) holds.
-    const intermediate = {
-        kind: 'operating-role-result',
-        schemaVersion: OPERATE_SCHEMA_VERSION,
-        protocolVersion: OPERATE_PROTOCOL_VERSION,
-        cycleId: input.cycleId,
-        roleId: input.mandate.roleId,
-        inputDigest: input.mandate.mandateDigest,
-        resultDigest: input.mandate.mandateDigest,
-        outcome: output.outcome,
-        proposals: output.proposals.map((proposal) => ({
-            proposalKey: proposal.proposalKey,
-            type: proposal.type,
-            title: proposal.title,
-            problem: proposal.problem,
-            proposal: proposal.proposal,
-            impact: proposal.impact,
-            confidence: proposal.confidence,
-            ease: proposal.ease,
-            severity: proposal.severity,
-            evidenceRefs: [],
-            ...(proposal.dependsOnProposalKeys
-                ? { dependsOnProposalKeys: proposal.dependsOnProposalKeys }
-                : {}),
-            ...(proposal.conflictsWithProposalKeys
-                ? { conflictsWithProposalKeys: proposal.conflictsWithProposalKeys }
-                : {}),
-            ...(proposal.sequenceProposalKeys
-                ? { sequenceProposalKeys: proposal.sequenceProposalKeys }
-                : {}),
-            citations: proposal.citations,
-        })),
-        gaps: output.gaps,
-        conflicts: output.conflicts,
-        producer: {
-            product: 'openplanr',
-            version: OPENPLANR_VERSION,
-            runtime: input.runtime,
-            capability,
-        },
-    };
-    // A quiet response has no proposals/citations, so it never touches the gate; a
-    // proposals response threads its citations through the already-live gate.
-    const gated = output.proposals.length > 0
-        ? await input.resolveCitations([intermediate])
-        : {
-            roleResults: [intermediate],
-            gaps: [],
-            notEvaluatedRoleIds: [],
-        };
-    const resolved = gated.roleResults[0] ?? intermediate;
-    // FR2: a proposals response whose citations resolved to ZERO evidence is
-    // not_evaluated — the gate returned this role in `notEvaluatedRoleIds` and
-    // opened the governed empty-grounding gap. The committed artifact is a
-    // schema-legal `quiet` result (no proposals); the not_evaluated state is
-    // carried by the gap and the returned flag.
-    const notEvaluated = gated.notEvaluatedRoleIds.includes(input.mandate.roleId);
-    // Finalize into a v1.2-valid committed operating-role-result: strip the
-    // now-resolved citations, keep the minted evidenceRefs, and let the surviving
-    // proposal count set the honest outcome (a not_evaluated response commits as
-    // quiet, its grounding failure preserved only as the opened gaps).
-    const survivingProposals = notEvaluated
-        ? []
-        : resolved.proposals
-            .map((proposal) => {
-            const { citations: _citations, ...rest } = proposal;
-            return rest;
-        })
-            .filter((proposal) => proposal.evidenceRefs.length > 0);
-    const unsigned = {
-        kind: 'operating-role-result',
-        schemaVersion: OPERATE_SCHEMA_VERSION,
-        protocolVersion: OPERATE_PROTOCOL_VERSION,
-        cycleId: input.cycleId,
-        roleId: input.mandate.roleId,
-        inputDigest: input.mandate.mandateDigest,
-        outcome: survivingProposals.length > 0 ? 'proposals' : 'quiet',
-        proposals: survivingProposals,
-        gaps: output.gaps,
-        conflicts: output.conflicts,
-        producer: {
-            product: 'openplanr',
-            version: OPENPLANR_VERSION,
-            runtime: input.runtime,
-            capability,
-        },
-    };
-    const result = {
-        ...unsigned,
-        resultDigest: protocol.computeOperatingRoleResultDigest(unsigned),
-    };
-    await assertOperatingArtifact('operating-role-result', result);
-    protocol.validateOperatingRoleResultDigest(result);
-    return { result, gaps: gated.gaps, notEvaluated };
-}
-function safeFailureMessage(error) {
-    try {
-        return sanitizeGeneratedPlainText(error instanceof Error ? error.message : String(error));
-    }
-    catch {
-        return 'Advisor failed with an unsafe or unredactable diagnostic.';
-    }
-}
-export function assertAdvisorIsolation(adapter) {
-    if (adapter.mode === 'native-isolated' &&
-        !['enforced', 'advisory', 'none'].includes(adapter.toolIsolation)) {
-        throw new OperateError('E_OPERATE_ADVISOR_ISOLATION', 'Native runtime advisors must declare their runtime-governed tool isolation.');
-    }
-    if (adapter.mode === 'structured' && adapter.toolIsolation !== 'not-applicable') {
-        throw new OperateError('E_OPERATE_ADVISOR_ISOLATION', 'Structured provider adapters must declare toolIsolation=not-applicable.');
-    }
-}
-export function createOfflineAdvisorAdapter(fixture) {
-    return {
-        id: 'offline-fixture',
-        mode: 'structured',
-        toolIsolation: 'not-applicable',
-        capability: 'analysis-standard',
-        parallelDispatch: false,
-        async invoke(input) {
-            if (fixture?.[input.roleId] !== undefined)
-                return fixture[input.roleId];
-            return input.mandate.protocolVersion === '1.4.0'
-                ? {
-                    outcome: 'quiet',
-                    analysisMarkdown: `# ${input.roleBrief.role.displayLabel}\n\nOffline fixture mode produced no recommendations.`,
-                    claims: [],
-                    actions: [],
-                    gaps: [],
-                    conflicts: [],
-                }
-                : { outcome: 'quiet', proposals: [], gaps: [], conflicts: [] };
-        },
-    };
-}
-class OpenPlanrStructuredAdapter {
-    id;
-    mode = 'structured';
-    toolIsolation = 'not-applicable';
-    capability = 'analysis-high';
-    parallelDispatch = false;
-    constructor(providerName) {
-        this.id = `openplanr-${providerName}`;
-    }
-    async invoke(input) {
-        void input;
-        throw new OperateError('E_OPERATE_PROVIDER_DEPRECATED', 'The structured-provider advisor path is deprecated; dispatch now runs through the native Protocol v1.3 mandate harness. See https://openplanr.dev/docs/operate/agent-harness. Scheduled for removal in OpenPlanr 2.0.0.');
-    }
-}
-/** Redacted error class label for diagnostics — no message or stack, ever. */
-function redactedProviderErrorClass(error) {
-    if (error instanceof AIError)
-        return `AIError:${error.code}`;
-    if (error instanceof Error && typeof error.name === 'string' && error.name.length > 0) {
-        return error.name;
-    }
-    return 'UnknownError';
-}
-/**
- * Build the actionable remedy for a failed structured-provider bootstrap,
- * preserving the underlying provider guidance (e.g. the `planr config set-key`
- * block an AIError already carries) and always naming the `--offline` escape.
- */
-function structuredProviderBootstrapRemedy(error, provider) {
-    const detail = error instanceof AIError
-        ? error.userMessage
-        : error instanceof Error
-            ? error.message
-            : String(error);
-    const trimmed = detail.trim();
-    const suffixParts = [];
-    if (!/config set-key/.test(trimmed)) {
-        suffixParts.push(`Configure a key with \`planr config set-key ${provider ?? '<provider>'}\``);
-    }
-    if (!/--offline/.test(trimmed)) {
-        suffixParts.push('or run the cycle offline with --offline');
-    }
-    const suffix = suffixParts.length > 0 ? ` ${suffixParts.join(' ')}.` : '';
-    return `Structured AI provider bootstrap failed: ${trimmed}${suffix}`;
-}
-export async function createConfiguredStructuredAdapter(projectRoot, options = {}) {
-    // `planr operate init` writes .planr/operate/config.json, not the project-wide
-    // .planr/config.json that loadConfig requires. A project that ran only the
-    // operate initializer therefore reaches this with no OpenPlanr config at all,
-    // and loadConfig throws ConfigNotFoundError — an untyped failure that surfaces
-    // as "an unexpected internal Operating Board error" on the primary first-run
-    // path. Both the missing config and the unconfigured provider mean the same
-    // thing to the operator, so both resolve to the same actionable error.
-    const config = await loadConfig(projectRoot).catch(() => null);
-    if (!config || !isAIConfigured(config)) {
-        throw new OperateError('E_OPERATE_ADVISOR_FAILED', 'No structured AI provider is configured; use --offline or configure OpenPlanr AI.');
-    }
-    // A named provider whose key cannot be resolved in this (possibly sandboxed)
-    // subprocess environment makes getAIProvider throw a raw AIError. Left
-    // unguarded it reaches index.ts's failure() as E_OPERATE_INTERNAL — the exact
-    // masked crash the audit reproduced. Convert any provider-bootstrap failure
-    // into a typed E_OPERATE_ADVISOR_FAILED that preserves the actionable remedy
-    // (`planr config set-key …` / `--offline`) and records a redacted error class.
-    let provider;
-    try {
-        provider = await getAIProvider(config, {
-            surface: 'operate-structured-provider',
-        });
-    }
-    catch (error) {
-        throw new OperateError('E_OPERATE_ADVISOR_FAILED', structuredProviderBootstrapRemedy(error, config.ai?.provider), { errorClass: redactedProviderErrorClass(error) });
-    }
-    void provider;
-    void options;
-    return new OpenPlanrStructuredAdapter(config.ai?.provider ?? 'ai');
-}
-export async function dispatchOperatingAdvisors(input) {
-    assertAdvisorIsolation(input.adapter);
-    const protocol = await loadOperatingProtocol();
-    const roleRegistry = protocol.listOperatingRoles();
-    // Canonical registry order so results and provenance are byte-identical across
-    // parallel/sequential dispatch and across the order roles arrive in (FR4).
-    const roleOrder = new Map(roleRegistry.map((role, index) => [role.id, index]));
-    // The runtime's isolation level is recorded once per dispatch. Protocol v1.4
-    // permits compatible native-agent workflows under runtime-governed session
-    // permissions; citation and schema validation still gate persistence.
-    const runtimeEnforcesBoundedReadOnly = await operatingRuntimeEnforcesBoundedReadOnly(input.runtime);
-    // A structured adapter cannot host a native lens. A native-isolated adapter
-    // can host either an enforced or runtime-governed native workflow.
-    const adapterNativeCapable = input.adapter.mode === 'native-isolated';
-    const runtimeWorkflowCapable = adapterNativeCapable;
-    const resolveIsolation = (roleId) => resolveOperatingDispatchIsolation({
-        roleId,
-        runtimeEnforcesBoundedReadOnly,
-        adapterNativeCapable,
-        runtimeWorkflowCapable,
-    });
-    // FR1/FR2: the mandate's declared read roots are the whole granted workspace —
-    // including a gitignored `.planr/` — never an evidence-index subset. Derived
-    // once and shared by every role's mandate on this path.
-    const roots = await deriveOperatingMandateRoots(input.projectRoot);
-    const skipped = [];
-    const runnable = [];
-    for (const role of input.readiness.roles) {
-        if (!role.modelCallAllowed || role.readiness === 'not_evaluated') {
-            skipped.push({
-                roleId: role.roleId,
-                gapId: role.gapId,
-                reason: role.missingEvidence.join('; '),
-            });
-            continue;
-        }
-        runnable.push(role);
-    }
-    // FR12: build the ONE shared, citation-bearing bootstrap map for this cycle —
-    // only when there is work to dispatch — and thread it, with the graceful per-role
-    // research budget, into every role's mandate below. Built once here and referenced
-    // by all runnable roles, never rebuilt per role, so five lenses stop
-    // independently re-walking the same planning and git indexes.
-    const researchBudgetMs = input.researchBudgetMs ?? DEFAULT_OPERATING_ROLE_RESEARCH_BUDGET_MS;
-    const bootstrapMap = runnable.length > 0 ? await buildOperatingBootstrapMap(input.projectRoot) : null;
-    async function dispatchRole(role) {
-        // Resolve THIS role's dispatch mode once and derive provenance from what is
-        // actually dispatched below — never re-derived after the fact. A role that
-        // resolves to a native bounded lens has its read-only tool grant enforced
-        // before the lens runs (below); `isolation` can only read
-        // `enforced-read-only-bounded` when the bounded grant was genuinely enforced.
-        const resolution = resolveIsolation(role.roleId);
-        const dispatch = {
-            isolation: resolution.isolation,
-            reconciliation: resolution.reconciliation,
-        };
-        let mandate;
-        try {
-            // FR1: a body-free operating mandate — the lens question, declared read
-            // boundaries, and citation requirement — replaces the curated evidence pack.
-            // FR12: the shared bootstrap map and graceful per-role research budget are
-            // threaded in as targeting guidance layered over the immutable mandate.
-            mandate = await buildOperatingMandate({
-                roleId: role.roleId,
-                roots,
-                runtime: input.runtime ?? input.adapter.id,
-                protocolVersion: input.protocolVersion ?? '1.3.0',
-                ...(bootstrapMap ? { bootstrapMap } : {}),
-                researchBudgetMs,
-            });
-        }
-        catch (error) {
-            return {
-                ok: false,
-                roleId: role.roleId,
-                message: safeFailureMessage(error),
-                modelCalls: 0,
-                dispatch,
-            };
-        }
-        const roleBrief = createRegistryReconciledAdvisorBrief(protocol, role.roleId);
-        const context = roleContext(input.context, role.roleId, new Set());
-        if (resolution.native) {
-            // Confine the native lens to the bounded read-only toolset over the
-            // mandate's declared roots. Constructing the toolset is what makes
-            // `enforced-read-only-bounded` true; a callable outside the read-only grant
-            // simply does not exist on the surface it hands the lens.
-            const grantedRoots = narrowMissionRootsToCeiling({
-                declaredRoots: mandate.boundaries.roots,
-                forbiddenPaths: mandate.boundaries.forbiddenPaths,
-            });
-            const toolset = createMissionToolset({
-                roots: grantedRoots,
-                ceiling: mandate.boundaries.sensitivityCeiling,
-            });
-            const grantedTools = Object.keys(toolset);
-            if (grantedTools.some((tool) => !MISSION_READ_ONLY_TOOLS.includes(tool))) {
-                throw new OperateError('E_OPERATE_PROVIDER_READ_ONLY', `Mission dispatch for ${role.roleId} assembled a tool outside the bounded read-only grant.`);
-            }
-        }
-        let built;
-        let lastError;
-        let roleModelCalls = 0;
-        for (let attempt = 0; attempt < 2; attempt += 1) {
-            try {
-                roleModelCalls += 1;
-                const response = await input.adapter.invoke({
-                    roleId: role.roleId,
-                    roleBrief,
-                    mandate,
-                    pinnedRevision: input.pinnedRevision,
-                    context,
-                    inputDigest: mandate.mandateDigest,
-                });
-                // The mandate response's citations flow into the injected universal gate,
-                // which mints the evidenceRefs that make the committed result v1.2-valid;
-                // a response that grounds zero evidence commits not_evaluated (quiet) with
-                // a governed gap. Identical finalization to the native adapter record path.
-                built = await createNativeMissionOperatingRoleResult({
-                    mandate,
-                    cycleId: input.cycleId,
-                    response,
-                    runtime: input.runtime ?? input.adapter.id,
-                    pinnedRevision: input.pinnedRevision,
-                    resolveCitations: input.resolveCitations,
-                });
-                break;
-            }
-            catch (error) {
-                lastError = error;
-            }
-        }
-        if (!built) {
-            return {
-                ok: false,
-                roleId: role.roleId,
-                message: safeFailureMessage(lastError),
-                modelCalls: roleModelCalls,
-                dispatch,
-            };
-        }
-        return {
-            ok: true,
-            result: built.result,
-            gaps: built.gaps,
-            modelCalls: roleModelCalls,
-            dispatch,
-        };
-    }
-    // Fan the per-role dispatch out in parallel where the adapter reports it,
-    // sequentially otherwise. FR12: the parallel path is bounded-concurrency so five
-    // lenses do not all research at once, without limiting any single lens's depth.
-    // Results are sorted into canonical registry order below so the reduced events
-    // are byte-identical across parallel and sequential dispatch and across dispatch
-    // order (FR4/E-004).
-    const dispatched = await runMissionDispatchFanOut({
-        items: runnable,
-        parallel: Boolean(input.adapter.parallelDispatch),
-        concurrency: input.concurrency ?? DEFAULT_OPERATING_DISPATCH_CONCURRENCY,
-        run: (role) => dispatchRole(role),
-    });
-    const dispatchByRole = new Map();
-    for (const entry of dispatched) {
-        dispatchByRole.set(entry.ok ? entry.result.roleId : entry.roleId, entry.dispatch);
-    }
-    const okEntries = dispatched.filter((entry) => entry.ok);
-    const results = okEntries
-        .map((entry) => entry.result)
-        .sort((left, right) => (roleOrder.get(left.roleId) ?? Number.MAX_SAFE_INTEGER) -
-        (roleOrder.get(right.roleId) ?? Number.MAX_SAFE_INTEGER));
-    // The governed gaps opened while resolving each role's citations (unresolvable
-    // citations, empty grounding) are threaded back so the engine records them
-    // alongside the cycle's other gaps.
-    const gaps = okEntries.flatMap((entry) => entry.gaps);
-    const failed = dispatched
-        .filter((entry) => !entry.ok)
-        .map(({ roleId, message }) => ({ roleId, message }));
-    const modelCalls = dispatched.reduce((total, entry) => total + entry.modelCalls, 0);
-    return {
-        results,
-        provenance: results.map((result) => {
-            const dispatchProvenance = dispatchByRole.get(result.roleId) ?? {
-                isolation: resolveIsolation(result.roleId).isolation,
-                reconciliation: resolveIsolation(result.roleId).reconciliation,
-            };
-            return {
-                roleId: result.roleId,
-                runtime: input.runtime ?? input.adapter.id,
-                adapterId: input.adapter.id,
-                capability: input.adapter.capability,
-                dispatch: input.adapter.parallelDispatch ? 'parallel' : 'sequential',
-                isolation: dispatchProvenance.isolation,
-                reconciliation: dispatchProvenance.reconciliation,
-            };
-        }),
-        skipped,
-        failed,
-        gaps,
-        blocked: failed.some((entry) => entry.roleId === 'chair') ||
-            (input.depth === 'deep' && failed.length > 0),
-        modelCalls,
-    };
-}
-/**
- * Standard-cycle advisor failures are governed data gaps, not ephemeral
- * warnings. IDs are temporary and are remapped against canonical state by the
- * cycle engine before persistence.
- */
-export async function advisorFailureGaps(input) {
-    const readinessByRole = new Map(input.readiness.roles.map((role) => [role.roleId, role]));
-    const gaps = [...input.failed]
-        .sort((left, right) => left.roleId.localeCompare(right.roleId))
-        .map(({ roleId, message }, index) => {
-        const role = readinessByRole.get(roleId);
-        return {
-            kind: 'operating-data-gap',
-            schemaVersion: OPERATE_SCHEMA_VERSION,
-            protocolVersion: OPERATE_PROTOCOL_VERSION,
-            id: `GAP-${String(index + 1).padStart(3, '0')}`,
-            cycleId: input.cycleId,
-            question: `What recovery or verified evidence is required to evaluate ${roleId}?`,
-            reason: `Advisor ${roleId} failed after its bounded retry: ${safeFailureMessage(message)}`,
-            unblocks: [],
-            affectedRoles: [roleId],
-            status: 'open',
-            owner: input.owner,
-            evidenceRefs: [...new Set(role?.evidenceRefs ?? [])].sort(),
-            createdAt: input.now,
-            updatedAt: input.now,
-        };
-    });
-    await Promise.all(gaps.map((gap) => assertOperatingArtifact('operating-data-gap', gap)));
-    return gaps;
-}
-export async function readProviderConsent(projectRoot, providerId) {
-    return readFile(path.join(projectRoot, '.planr', 'operate', 'providers', `${providerId}.json`), 'utf8')
-        .then(async (raw) => {
-        const manifest = JSON.parse(raw);
-        await assertOperatingArtifact('operating-provider-manifest', manifest);
-        (await loadOperatingProtocol()).validateOperatingProviderPolicyDigest(manifest);
-        return manifest;
-    })
-        .catch(() => null);
-}
-function providerPolicyCandidate(input, acceptedAt, existing) {
-    return {
-        kind: 'operating-provider-manifest',
-        schemaVersion: OPERATE_SCHEMA_VERSION,
-        protocolVersion: OPERATE_PROTOCOL_VERSION,
-        id: input.id,
-        providerId: input.providerId,
-        providerVersion: input.providerVersion,
-        mode: input.mode,
-        readOnly: true,
-        endpoint: structuredClone(input.endpoint),
-        permittedDataClasses: [...new Set(input.permittedDataClasses)].sort(),
-        retention: structuredClone(input.retention),
-        capabilities: {
-            incremental: input.incremental,
-            deep: input.deep,
-            toolIsolation: input.toolIsolation,
-        },
-        limits: structuredClone(input.limits),
-        consent: {
-            policyVersion: input.consentPolicyVersion,
-            status: existing ? 'renewed' : 'first-use',
-            acceptedAt: existing?.consent.acceptedAt ?? acceptedAt,
-            renewedAt: existing ? acceptedAt : null,
-            nextReviewAt: new Date(Date.parse(acceptedAt) + 180 * 24 * 60 * 60 * 1_000).toISOString(),
-            renewalTriggers: [...new Set(input.renewalTriggers)].sort(),
-        },
-        policyDigest: `sha256:${'0'.repeat(64)}`,
-        configurationDigest: input.configurationDigest,
-        capturedAt: acceptedAt,
-    };
-}
-/**
- * Build and, only after explicit authority, persist the credential-free policy
- * that controls a remote advisor call. Keys and endpoint secrets never enter
- * the manifest.
- */
-export async function ensureOperatingProviderConsent(input) {
-    const existing = await readProviderConsent(input.projectRoot, input.provider.providerId);
-    const now = input.now ?? new Date().toISOString();
-    const protocol = await loadOperatingProtocol();
-    const candidate = providerPolicyCandidate(input.provider, now, existing);
-    candidate.policyDigest = protocol.computeOperatingProviderPolicyDigest(candidate);
-    await assertOperatingArtifact('operating-provider-manifest', candidate);
-    protocol.validateOperatingProviderPolicyDigest(candidate);
-    const changed = !existing ||
-        existing.policyDigest !== candidate.policyDigest ||
-        existing.configurationDigest !== candidate.configurationDigest ||
-        Date.parse(existing.consent.nextReviewAt ?? '1970-01-01') <= Date.parse(now);
-    if (!changed)
-        return { manifest: existing, changed: false };
-    if (!input.confirmed) {
-        throw new OperateError('E_OPERATE_AUTHORITY_REQUIRED', 'The disclosed provider policy requires explicit first-use or renewal consent.', {
-            provider: candidate.providerId,
-            endpoint: candidate.endpoint,
-            permittedDataClasses: candidate.permittedDataClasses,
-            retention: candidate.retention,
-            limits: candidate.limits,
-            policyDigest: candidate.policyDigest,
-        });
-    }
-    if (input.persist !== false) {
-        const directory = path.join(input.projectRoot, '.planr', 'operate', 'providers');
-        await mkdir(directory, { recursive: true, mode: 0o700 });
-        const target = path.join(directory, `${candidate.providerId}.json`);
-        const temporary = `${target}.${process.pid}.${randomUUID()}.tmp`;
-        await writeFile(temporary, `${canonicalize(candidate)}\n`, { mode: 0o600 });
-        await rename(temporary, target);
-    }
-    return { manifest: candidate, changed: true };
-}
-function providerEndpoint(config) {
-    const provider = config.ai?.provider ?? 'openai';
-    const configuredUrl = config.ai?.ollamaBaseUrl?.trim();
-    const defaultUrl = provider === 'anthropic'
-        ? 'https://api.anthropic.com'
-        : provider === 'openai'
-            ? 'https://api.openai.com'
-            : 'http://127.0.0.1:11434';
-    const identity = configuredUrl || defaultUrl;
-    let display = provider === 'ollama' ? 'local Ollama' : `${provider} API`;
-    try {
-        const parsed = new URL(identity);
-        display = `${provider} @ ${parsed.origin}`;
-    }
-    catch {
-        display = `${provider} @ configured endpoint`;
-    }
-    return {
-        kind: provider === 'ollama' ? 'local' : 'remote',
-        display,
-        authentication: provider === 'ollama' ? 'none' : 'machine-local',
-        identity,
-    };
-}
-export function configuredAdvisorProviderPolicy(input) {
-    const provider = input.config.ai?.provider ?? 'openai';
-    const model = input.config.ai?.model ?? DEFAULT_MODELS[provider];
-    const endpoint = providerEndpoint(input.config);
-    const local = endpoint.kind === 'local';
-    const configurationDigest = canonicalDigest({
-        provider,
-        model,
-        endpoint: endpoint.identity,
-        runtime: input.runtime,
-        adapterId: input.adapterId,
-        permittedDataClasses: [
-            'source-code',
-            'planning-artifacts',
-            'git-metadata',
-            'issue-metadata',
-            'project-metadata',
-        ],
-        retention: local
-            ? { providerStoresRequestContent: false, maxProviderRetentionDays: 0 }
-            : { providerStoresRequestContent: true, maxProviderRetentionDays: 30 },
-    });
-    return {
-        id: `PRV-openplanr-${provider}`,
-        providerId: `openplanr-${provider}`,
-        providerVersion: '1.0.0',
-        mode: 'structured',
-        toolIsolation: 'not-applicable',
-        endpoint: {
-            kind: endpoint.kind,
-            display: `${endpoint.display} · ${model}`,
-            authentication: endpoint.authentication,
-            redacted: true,
-        },
-        permittedDataClasses: [
-            'source-code',
-            'planning-artifacts',
-            'git-metadata',
-            'issue-metadata',
-            'project-metadata',
-        ],
-        retention: {
-            providerStoresRequestContent: !local,
-            maxProviderRetentionDays: local ? 0 : 30,
-            localEvidenceRetention: 'cycle',
-        },
-        incremental: false,
-        deep: true,
-        limits: {
-            maxItems: 2_000,
-            maxBytes: 10 * 1024 * 1024,
-            maxDurationMs: 60_000,
-            maxRequests: 12,
-            maxTokens: 49_152,
-            maxCostUsd: null,
-        },
-        consentPolicyVersion: '1.0.0',
-        renewalTriggers: ['policy-change', 'scope-expansion', 'credential-renewal', 'scheduled-review'],
-        configurationDigest,
-    };
-}
-//# sourceMappingURL=advisors.js.map