[{"_1":2,"_3":-5,"_4":-5},"loaderData",{"_5":6,"_7":8,"_9":10},"actionData","errors","root",{"_265":266,"_267":268},"routes/_app.x.$handle.$name",{"_97":96,"_31":44,"_106":85,"_26":107,"_13":14,"_108":-5,"_11":109,"_110":111,"_112":113,"_114":115,"_116":117,"_20":-5,"_118":-5,"_119":120,"_121":24,"_122":123,"_124":125,"_126":-5,"_127":-5,"_128":129},"routes/_app.x.$handle.$name.branch.$branch.version.$version.diff",{"_11":12,"_13":14,"_15":16,"_17":-5,"_18":-5},"project",{"_42":43,"_31":44,"_37":45,"_46":-5,"_47":24,"_48":49,"_50":49,"_51":52,"_53":-5,"_18":54,"_55":56,"_57":24,"_58":-5,"_59":-5,"_60":61,"_62":63,"_64":-5,"_65":-5,"_66":-5,"_67":-5,"_68":69,"_70":69,"_71":72,"_73":72,"_74":72,"_75":69,"_76":69,"_77":78,"_79":80,"_81":82,"_83":24},"version",8,"changes",[19],"aiChatInfo","userId",{"_20":21,"_22":-7,"_23":24,"_25":24,"_26":27,"_28":27},"file",{"_29":30,"_31":32,"_33":34,"_35":36,"_37":38,"_39":14,"_40":41,"_23":24},"previous","deleted",false,"deletedHere","path","vendor/rumbo-core/src/revision.js","latestPath","fileId","7b1297d6-af4d-11f1-9522-1607ee4eb77e","name","revision.js","fileType","script","content","import { stableJson, stableValue } from './canonical.js';\n\nexport function normalizeRevision(binding) {\n let canonical;\n try {\n canonical = stableValue(binding);\n } catch {\n throw new Error('INVALID_REVISION');\n }\n if (!canonical || typeof canonical !== 'object' || Array.isArray(canonical)) {\n throw new Error('INVALID_REVISION');\n }\n const keys = Object.keys(canonical);\n if (keys.some((key) => !['kind', 'value', 'scope'].includes(key))) throw new Error('INVALID_REVISION');\n\n const kind = typeof canonical.kind === 'string' ? canonical.kind.trim() : '';\n const value = typeof canonical.value === 'string' ? canonical.value.trim() : '';\n if (!kind || !value) throw new Error('INVALID_REVISION');\n\n const normalized = { kind, value };\n if ('scope' in canonical) {\n const scope = typeof canonical.scope === 'string' ? canonical.scope.trim() : '';\n if (!scope) throw new Error('INVALID_REVISION');\n normalized.scope = scope;\n }\n return normalized;\n}\nexport function revisionMatches(expected, observed) {\n try {\n return stableJson(normalizeRevision(expected)) === stableJson(normalizeRevision(observed));\n } catch {\n return false;\n }\n}\n","createdAt",["D",1789288273685],"branchVersion","branchId","3dfd20c4-436f-4774-915d-073b4bed3530","id","1991cca8-bf13-4945-bf68-f1df4a1a5770","rumbo-agent-reliability-mcp",["D",1788582330391],"deletedAt","legacyEnvVars","privacy","public","httpPrivacy","description","RUMBO agent reliability MCP tools","imageUrl","d3788f19-13c5-4356-87e8-dce639626bb9","pinnedFileId","01a06fd0-810d-7419-8d90-6ae071d33f35","pinPreview","pinMigrated","screenshotVersion","likeCount",0,"orgId","01a04116-f28c-754c-9173-77bf5814fcd6","tursoName","tursoHostname","tursoMeasuredAt","pinned","userPerm","r","valPerm","blobPerm","rw","emailPerm","sqlitePerm","projectPerm","telemetryPerm","user",{"_97":96,"_105":101},"org",{"_97":96,"_98":99,"_100":101,"_18":54,"_102":103,"_104":24},"branch",{"_13":84,"_31":85,"_42":41,"_86":43,"_87":88,"_89":90,"_91":49,"_92":43,"_93":63,"_94":44,"_95":96},"isOwned",32,"cea-v1-2-candidate","projectId","forkedBranchId","ebecba93-11f3-4e98-9277-46a8ff8ddda6","forkedBranchVersion",15,"forkedProjectPrivacy","forkedProjectId","forkedProjectOrgId","projectName","userHandle","sebas1","handle","displayName","sebas","avatar","https://lh3.googleusercontent.com/a/ACg8ocJaC6uiXPmN9dvgA2XcSlzBSkPJO0GN5zKzPm3a8cbjmXC9HYA=s96-c","tier","free","canRestrictVals","profileImageUrl","branchName","","adoptOrgId",{"_42":43,"_31":44,"_37":261,"_46":-5,"_47":24,"_48":49,"_50":49,"_51":52,"_53":-5,"_18":54,"_55":56,"_57":24,"_58":-5,"_59":-5,"_60":61,"_62":63,"_64":-5,"_65":-5,"_66":-5,"_67":-5,"_68":69,"_70":69,"_71":72,"_73":72,"_74":72,"_75":69,"_76":69,"_77":262,"_79":263,"_81":264,"_83":24},"tree",[197,203,204,205,206,207],"files",[197,198,199,200,201,202,203,204,205,206,207],"counts",{"_186":61,"_187":188,"_189":61,"_190":191,"_192":193},"contents",[136,137,138,139,140,141],"upstreamProject","upstreamBranch",{"_42":88,"_37":130,"_131":132,"_31":133,"_86":43,"_13":90,"_87":-5,"_89":-5,"_134":135,"_46":-5,"_83":24,"_48":49,"_95":96,"_62":63,"_94":44},"emptyEnvVars","flags",[],"hasHttpFiles",true,"grants","orgs","customDomains",{},["D",1788582330391],"updatedAt",["D",1789289681119],"main","lockfileKey","1789289681251-esm.town_v_sebas1_rumbo-agent-reliability-mcp@15-main",{"_42":179,"_31":180,"_29":181,"_145":-5,"_40":41,"_39":182,"_147":125,"_37":183,"_13":61,"_33":184,"_35":-5,"_23":24,"_150":-5,"_26":185},{"_42":173,"_31":174,"_29":175,"_145":-5,"_40":41,"_39":146,"_147":125,"_37":176,"_13":61,"_33":20,"_35":177,"_23":24,"_150":-5,"_26":178},{"_42":166,"_31":167,"_29":168,"_145":-5,"_40":41,"_39":169,"_147":24,"_37":170,"_13":61,"_33":20,"_35":171,"_23":24,"_150":-5,"_26":172},{"_42":159,"_31":160,"_29":161,"_145":-5,"_40":41,"_39":146,"_147":24,"_37":162,"_13":61,"_33":163,"_35":164,"_23":24,"_150":-5,"_26":165},{"_42":152,"_31":153,"_29":154,"_145":-5,"_40":41,"_39":146,"_147":24,"_37":155,"_13":61,"_33":156,"_35":157,"_23":24,"_150":-5,"_26":158},{"_42":142,"_31":143,"_29":144,"_145":-5,"_40":41,"_39":146,"_147":125,"_37":148,"_13":61,"_33":20,"_35":149,"_23":24,"_150":-5,"_26":151},"3a0a09c6-2939-4f37-9377-5882da5c02ca","rumbo-semantic-core.mjs","01a099e2-134e-77a7-a880-2796bf303c5b","parentId",1,"latest",["D",1789288125263],"/**\n * RUMBO Agent Reliability semantic hardening core v1.3.12\n * Dependency-free, deterministic, fail-closed.\n * Designed to sit behind the existing MCP tool names without changing their input schemas.\n */\n\nconst IDS = Object.freeze({\n csr: 'RUMBO-CSR-V1.3.12',\n drr: 'RUMBO-DRR-V1.3.12',\n glc: 'RUMBO-GLC-V1.3.12',\n evc: 'RUMBO-EVC-V1.3.12',\n afs: 'RUMBO-AFS-V1.3.12',\n});\n\nfunction asText(v) { return typeof v === 'string' ? v.trim() : ''; }\n\nfunction strictJsonParse(text) {\n const s = String(text ?? '').trim();\n let i = 0;\n const fail = (code) => { throw new Error(code); };\n const ws = () => { while (i < s.length && /\\s/.test(s[i])) i += 1; };\n\n const parseStringToken = () => {\n if (s[i] !== '\"') fail('INVALID_JSON_STRING');\n const start = i;\n i += 1;\n while (i < s.length) {\n const ch = s[i];\n if (ch === '\"') {\n i += 1;\n try { return JSON.parse(s.slice(start, i)); } catch { fail('INVALID_JSON_STRING'); }\n }\n if (ch === '\\\\') {\n if (i + 1 >= s.length) fail('INVALID_JSON_ESCAPE');\n if (s[i + 1] === 'u') {\n if (!/^[0-9A-Fa-f]{4}$/.test(s.slice(i + 2, i + 6))) fail('INVALID_JSON_UNICODE_ESCAPE');\n i += 6;\n } else {\n if (!/[\"\\\\/bfnrt]/.test(s[i + 1])) fail('INVALID_JSON_ESCAPE');\n i += 2;\n }\n continue;\n }\n if (ch.charCodeAt(0) < 0x20) fail('INVALID_JSON_CONTROL_CHAR');\n i += 1;\n }\n fail('UNTERMINATED_JSON_STRING');\n };\n\n const parseValue = () => {\n ws();\n if (i >= s.length) fail('UNEXPECTED_JSON_EOF');\n if (s[i] === '{') return parseObject();\n if (s[i] === '[') return parseArray();\n if (s[i] === '\"') { parseStringToken(); return; }\n const start = i;\n while (i < s.length && !/[\\s,\\]}]/.test(s[i])) i += 1;\n const token = s.slice(start, i);\n if (!token) fail('INVALID_JSON_VALUE');\n try { JSON.parse(token); } catch { fail('INVALID_JSON_VALUE'); }\n };\n\n const parseObject = () => {\n i += 1;\n ws();\n const seen = new Set();\n if (s[i] === '}') { i += 1; return; }\n while (i < s.length) {\n ws();\n const key = parseStringToken();\n if (seen.has(key)) fail(`DUPLICATE_JSON_KEY:${key}`);\n seen.add(key);\n ws();\n if (s[i] !== ':') fail('INVALID_JSON_OBJECT_COLON');\n i += 1;\n parseValue();\n ws();\n if (s[i] === '}') { i += 1; return; }\n if (s[i] !== ',') fail('INVALID_JSON_OBJECT_SEPARATOR');\n i += 1;\n }\n fail('UNTERMINATED_JSON_OBJECT');\n };\n\n const parseArray = () => {\n i += 1;\n ws();\n if (s[i] === ']') { i += 1; return; }\n while (i < s.length) {\n parseValue();\n ws();\n if (s[i] === ']') { i += 1; return; }\n if (s[i] !== ',') fail('INVALID_JSON_ARRAY_SEPARATOR');\n i += 1;\n }\n fail('UNTERMINATED_JSON_ARRAY');\n };\n\n parseValue();\n ws();\n if (i !== s.length) fail('TRAILING_JSON_DATA');\n return JSON.parse(s);\n}\n\nfunction jsonInputState(v) {\n if (v && typeof v === 'object') return { kind: 'VALID', value: v, error: null };\n if (typeof v !== 'string') return { kind: 'NOT_JSON', value: null, error: null };\n const s = v.trim();\n if (!s || !['{', '['].includes(s[0])) return { kind: 'NOT_JSON', value: null, error: null };\n try { return { kind: 'VALID', value: strictJsonParse(s), error: null }; }\n catch (err) { return { kind: 'INVALID', value: null, error: String(err?.message || 'INVALID_JSON') }; }\n}\n\nfunction tryJson(v) {\n const state = jsonInputState(v);\n return state.kind === 'VALID' ? state.value : null;\n}\n\nfunction cleanToken(v) {\n return String(v ?? '').trim().replace(/[.;,!?:]+$/g, '').trim();\n}\n\nfunction extractAssignments(text) {\n const out = [];\n const re = /\\b([A-Za-zÁÉÍÓÚÜÑáéíóúüñ_][\\wÁÉÍÓÚÜÑáéíóúüñ-]*)\\s*=\\s*([^\\s,;]+)/g;\n for (const m of asText(text).matchAll(re)) out.push({ key: m[1], value: cleanToken(m[2]), raw: m[0] });\n return out;\n}\n\nfunction flattenJsonScalarAssignments(value, prefix = '') {\n const out = [];\n if (value === null || typeof value !== 'object') {\n if (prefix) out.push({ key: prefix, value: cleanToken(value), raw: `${prefix}=${String(value)}` });\n return out;\n }\n if (Array.isArray(value)) {\n value.forEach((item, index) => out.push(...flattenJsonScalarAssignments(item, prefix ? `${prefix}[${index}]` : `[${index}]`)));\n return out;\n }\n for (const [key, item] of Object.entries(value)) {\n const path = prefix ? `${prefix}.${key}` : key;\n out.push(...flattenJsonScalarAssignments(item, path));\n }\n return out;\n}\n\nfunction semanticAssignments(text) {\n const state = jsonInputState(text);\n if (state.kind === 'VALID') return flattenJsonScalarAssignments(state.value);\n return extractAssignments(text);\n}\n\nfunction lastAssignment(text, preferredKey) {\n const a = extractAssignments(text).filter(x => !preferredKey || x.key === preferredKey);\n return a.length ? a[a.length - 1] : null;\n}\n\nconst NEGATIVE_SCOPE_RE = /\\b(?:SUPERSEDED|HISTORICAL|ARCHIVED|REJECTED|NO_AUTHORITY|SUPERSEDIDO|HIST[OÓ]RICO|ARCHIVADO|RECHAZADO)\\b/i;\nconst SELECTED_SCOPE_RE = /\\b(?:SELECTED|CANONICAL_SELECTED|SELECCIONADO|CANDIDATO_SELECCIONADO)\\b/i;\n\nfunction scopedAssignments(text) {\n const input = asText(text);\n const out = [];\n const re = /\\b([A-Za-zÁÉÍÓÚÜÑáéíóúüñ_][\\wÁÉÍÓÚÜÑáéíóúüñ-]*)\\s*=\\s*([^\\s,;]+)/g;\n for (const m of input.matchAll(re)) {\n const left = input.slice(Math.max(0, m.index - 120), m.index);\n const segment = left.split(/[;\\n.]/).at(-1) ?? left;\n out.push({\n key: m[1],\n value: cleanToken(m[2]),\n raw: m[0],\n selected: SELECTED_SCOPE_RE.test(segment),\n excluded: NEGATIVE_SCOPE_RE.test(segment),\n });\n }\n return out;\n}\n\nconst POSITIVE_EVIDENCE_STATUS = /^(?:PASS|TRUE|OK|VERIFIED)$/i;\nconst NEGATIVE_EVIDENCE_STATUS = /^(?:FAIL|FAILED|FALSE|UNVERIFIED|NO_GO|NOT_PROVEN|UNPROVEN)$/i;\n\nfunction evidenceCategoryState(text, key) {\n const input = asText(text);\n const re = new RegExp(`\\\\b${key}\\\\s*=\\\\s*([^\\\\s,;]+)`, 'ig');\n const rawValues = [...input.matchAll(re)].map(m => cleanToken(m[1]));\n if (!rawValues.length) return { present: false, positive: false, conflict: false, invalid: false, values: [] };\n\n const semantic = rawValues.map(v => {\n if (POSITIVE_EVIDENCE_STATUS.test(v)) return 'POSITIVE';\n if (NEGATIVE_EVIDENCE_STATUS.test(v)) return 'NEGATIVE';\n return 'INVALID';\n });\n const distinct = [...new Set(semantic)];\n const invalid = distinct.includes('INVALID');\n const conflict = distinct.includes('POSITIVE') && distinct.includes('NEGATIVE');\n const positive = !invalid && !conflict && distinct.length === 1 && distinct[0] === 'POSITIVE';\n return { present: true, positive, conflict, invalid, values: rawValues };\n}\n\nfunction externalGateCategory(text) {\n const input = asText(text);\n const re = /\\bexternal_gate_status\\s*=\\s*([^;\\n]+)/ig;\n const values = [...input.matchAll(re)].map(m => cleanToken(m[1])).filter(Boolean);\n if (!values.length) return { present: false, clear: false, conflict: false, values: [], value: null };\n\n const states = values.map(value => /^(?:NONE|CLEAR|PASS|NO_EXTERNAL_GATES?|NO_EXTERNAL_GATES_PENDING)$/i.test(value) ? 'CLEAR' : 'OPEN');\n const distinct = [...new Set(states)];\n const conflict = distinct.length > 1;\n const clear = !conflict && distinct.length === 1 && distinct[0] === 'CLEAR';\n return { present: true, clear, conflict, values, value: values.at(-1) ?? null };\n}\n\nfunction extractTarget(goal) {\n const json = tryJson(goal);\n if (json && typeof json === 'object') {\n if (json.target != null) return json.target;\n if (json.TARGET != null) return json.TARGET;\n }\n const t = asText(goal);\n const marker = t.match(/\\bTARGET(?:\\s+expl[ií]cito)?\\s*:\\s*([^\\n]+)/i);\n const fromMarker = marker ? lastAssignment(marker[1]) : null;\n if (fromMarker) return { [fromMarker.key]: fromMarker.value };\n const any = lastAssignment(t);\n return any ? { [any.key]: any.value } : null;\n}\n\nfunction currentFromEvidence(evidence, priorState, preferredKey = '') {\n const ej = tryJson(evidence);\n if (ej && typeof ej === 'object') {\n const selected = ej.selected ?? ej.SELECTED ?? ej.canonicalSelected ?? ej.canonical_selected ?? null;\n if (selected != null) {\n if (preferredKey && selected && typeof selected === 'object' && !Array.isArray(selected) && Object.hasOwn(selected, preferredKey)) {\n return { [preferredKey]: cleanToken(selected[preferredKey]) };\n }\n if (selected && typeof selected === 'object' && !Array.isArray(selected)) return selected;\n }\n const direct = ej.current ?? ej.CURRENT ?? ej.state ?? null;\n if (direct != null) {\n if (preferredKey && direct && typeof direct === 'object' && !Array.isArray(direct) && Object.hasOwn(direct, preferredKey)) {\n return { [preferredKey]: cleanToken(direct[preferredKey]) };\n }\n return direct;\n }\n }\n\n const text = asText(evidence);\n const scoped = scopedAssignments(text);\n const selectedScoped = scoped.filter(x => x.selected && !x.excluded && (!preferredKey || x.key === preferredKey));\n if (selectedScoped.length) {\n const selected = selectedScoped.at(-1);\n return { [selected.key]: selected.value };\n }\n\n const primary = text.match(/(?:EVIDENCIA\\s+PRIMARIA\\s+ACTUAL|CURRENT|ESTADO\\s+ACTUAL)\\s*:\\s*([^\\n.]+)/i);\n if (primary) {\n const primaryAssignments = scopedAssignments(primary[1]).filter(x => !x.excluded);\n const selected = preferredKey\n ? primaryAssignments.filter(x => x.key === preferredKey).at(-1)\n : primaryAssignments.at(-1);\n if (selected) return { [selected.key]: selected.value };\n }\n\n const evidenceAssignments = scoped.filter(x => !x.excluded);\n if (preferredKey) {\n const preferred = evidenceAssignments.filter(x => x.key === preferredKey).at(-1);\n if (preferred) return { [preferred.key]: preferred.value };\n }\n\n const priorAssignment = lastAssignment(priorState);\n if (priorAssignment) {\n const sameKeyCurrent = evidenceAssignments.filter(x => x.key === priorAssignment.key).at(-1);\n if (sameKeyCurrent) return { [sameKeyCurrent.key]: sameKeyCurrent.value };\n }\n\n // Unscoped current evidence outranks priorState. Explicitly historical/superseded\n // assignments are excluded before freshness selection.\n const freshestEvidence = evidenceAssignments.at(-1);\n if (freshestEvidence) return { [freshestEvidence.key]: freshestEvidence.value };\n\n if (priorAssignment) return { [priorAssignment.key]: priorAssignment.value };\n return null;\n}\n\nfunction equals(a, b) { return JSON.stringify(a) === JSON.stringify(b); }\n\nfunction resultEnvelope(capability, id, result, classification, confidence, evidenceGrounding) {\n // Keep the structured `result` while mirroring result fields at top level for\n // backward compatibility with the live V1.2 response shape.\n return { capability, id, ...result, result, classification, confidence, evidenceGrounding };\n}\n\nexport function canonicalStateRecovery({ goal, evidence, priorState = '' }) {\n const goalJsonState = jsonInputState(goal);\n const evidenceJsonState = jsonInputState(evidence);\n const invalidJson = [\n ...(goalJsonState.kind === 'INVALID' ? [{ field: 'goal', error: goalJsonState.error }] : []),\n ...(evidenceJsonState.kind === 'INVALID' ? [{ field: 'evidence', error: evidenceJsonState.error }] : []),\n ];\n if (invalidJson.length) {\n return resultEnvelope('RUMBO Canonical State Recovery', IDS.csr,\n { current: null, target: null, currentRecovered: false, targetSpecified: false, targetReached: false, nextGate: 'INVALID_JSON_INPUT', jsonErrors: invalidJson },\n 'FAIL_CLOSED_INVALID_JSON', 'unknown',\n { usedPriorState: Boolean(asText(priorState)), evidencePresent: Boolean(asText(evidence)), precedence: 'STRICT_JSON_BEFORE_SEMANTIC_RECOVERY', terminalPunctuationNormalized: true, selectedMarkerUsed: false, suppressedHistoricalAssignments: 0, strictJsonValid: false });\n }\n const target = extractTarget(goal);\n const targetKey = target && typeof target === 'object' && !Array.isArray(target) ? Object.keys(target)[0] : '';\n const current = currentFromEvidence(evidence, priorState, targetKey);\n const ej = tryJson(evidence);\n let nextGate = null;\n if (ej && typeof ej === 'object' && ej.nextGate) nextGate = ej.nextGate;\n if (!nextGate) {\n const text = asText(evidence);\n const explicit = text.match(/(?:[uú]nica|unica)\\s+transici[oó]n\\s+v[aá]lida(?:\\s+disponible)?\\s+(?:es|:)\\s*([A-Z][A-Z0-9_-]*)/)\n || text.match(/next(?:Gate|\\s+gate|\\s+transition)\\s*[:=]\\s*([A-Za-z0-9_-]+)/i);\n if (explicit) nextGate = explicit[1];\n }\n if (!nextGate && current && target && !equals(current, target)) {\n const ck = Object.keys(current)[0];\n const tk = Object.keys(target)[0];\n nextGate = ck && tk && ck === tk\n ? `TRANSITION ${ck}=${current[ck]} -> ${tk}=${target[tk]}`\n : 'APPLY_SINGLE_VALID_TRANSITION';\n }\n if (current && target && equals(current, target)) nextGate = 'NONE';\n const currentRecovered = Boolean(current);\n const targetSpecified = Boolean(target);\n const targetReached = Boolean(current && target && equals(current, target));\n const confidence = currentRecovered ? 'proven' : 'unknown';\n const classification = !currentRecovered\n ? 'FAIL_CLOSED_INSUFFICIENT_EVIDENCE'\n : targetSpecified\n ? 'RESOLVED'\n : 'CURRENT_RECOVERED_TARGET_UNSPECIFIED';\n const scoped = scopedAssignments(asText(evidence));\n const jsonEvidence = tryJson(evidence);\n const selectedMarkerUsed = scoped.some(x => x.selected && !x.excluded)\n || Boolean(jsonEvidence && typeof jsonEvidence === 'object' && (jsonEvidence.selected ?? jsonEvidence.SELECTED ?? jsonEvidence.canonicalSelected ?? jsonEvidence.canonical_selected));\n const suppressedHistoricalAssignments = scoped.filter(x => x.excluded).length;\n return resultEnvelope(\n 'RUMBO Canonical State Recovery', IDS.csr,\n { current, target, currentRecovered, targetSpecified, targetReached, nextGate: nextGate ?? 'EVIDENCE_REQUIRED' },\n classification, confidence,\n { usedPriorState: Boolean(asText(priorState)), evidencePresent: Boolean(asText(evidence)), precedence: 'SELECTED_THEN_CURRENT_EVIDENCE_THEN_PRIOR_STATE', terminalPunctuationNormalized: true, selectedMarkerUsed, suppressedHistoricalAssignments }\n );\n}\n\nexport function deepResearchReconcile({ question, evidence, constraints = '' }) {\n const evidenceJsonState = jsonInputState(evidence);\n if (evidenceJsonState.kind === 'INVALID') {\n return resultEnvelope('RUMBO Deep Research Reconcile', IDS.drr,\n { facts: [], conclusion: null, reconciledConclusion: null, conflicts: [{ key: 'JSON', values: [evidenceJsonState.error], basis: 'STRICT_JSON_REQUIRED' }], superseded: [], unresolved: ['INVALID_JSON_EVIDENCE'] },\n 'FAIL_CLOSED_INVALID_JSON', 'unknown', { question: asText(question), constraints: asText(constraints), strictJsonValid: false });\n }\n const ej = tryJson(evidence);\n if (ej && typeof ej === 'object') {\n const facts = Array.isArray(ej.facts) ? ej.facts : [];\n const conclusion = ej.reconciledConclusion ?? ej.conclusion ?? ej.current ?? ej.resolution ?? null;\n const unresolved = Array.isArray(ej.unresolved) ? ej.unresolved : (conclusion ? [] : ['No explicit resolution in evidence JSON']);\n return resultEnvelope('RUMBO Deep Research Reconcile', IDS.drr,\n { facts, conclusion, reconciledConclusion: conclusion, conflicts: ej.conflicts ?? [], superseded: ej.superseded ?? [], unresolved },\n conclusion ? 'RECONCILED' : 'FAIL_CLOSED_INSUFFICIENT_EVIDENCE', conclusion ? 'proven' : 'unknown',\n { question: asText(question), constraints: asText(constraints) });\n }\n\n const text = asText(evidence);\n const primary = text.match(/Fuente\\s+primaria\\s+actual\\s*:\\s*([\\wÁÉÍÓÚÜÑáéíóúüñ-]+\\s*=\\s*[^.\\s]+)/i);\n const secondary = text.match(/Fuente\\s+secundaria\\s+antigua\\s*:\\s*([\\wÁÉÍÓÚÜÑáéíóúüñ-]+\\s*=\\s*[^.\\s]+)/i);\n if (primary) {\n const conclusion = cleanToken(primary[1]);\n const superseded = secondary ? [cleanToken(secondary[1])] : [];\n const facts = [{ raw: conclusion, authority: 'primary/current' }];\n return resultEnvelope('RUMBO Deep Research Reconcile', IDS.drr,\n { facts, conclusion, reconciledConclusion: conclusion, conflicts: secondary ? [{ preferred: conclusion, superseded: cleanToken(secondary[1]), basis: 'primary/current authority' }] : [], superseded, unresolved: [] },\n 'RECONCILED', 'proven', { authorityRule: 'primary/current over secondary/older' });\n }\n\n const assignments = extractAssignments(text);\n if (assignments.length) {\n const byKey = new Map();\n for (const item of assignments) {\n if (!byKey.has(item.key)) byKey.set(item.key, []);\n byKey.get(item.key).push(item);\n }\n const facts = [];\n const conflicts = [];\n const superseded = [];\n for (const [key, items] of byKey) {\n const latest = items.at(-1);\n facts.push({ key, value: latest.value, raw: latest.raw });\n const unique = [...new Set(items.map(x => x.value))];\n if (unique.length > 1) {\n conflicts.push({ key, values: unique, preferred: latest.value, basis: 'freshest supplied evidence order' });\n superseded.push(...items.slice(0, -1).map(x => `${x.key}=${x.value}`));\n }\n }\n const reconciledConclusion = facts.map(x => `${x.key}=${x.value}`).join('; ');\n return resultEnvelope('RUMBO Deep Research Reconcile', IDS.drr,\n { facts, conclusion: reconciledConclusion, reconciledConclusion, conflicts, superseded, unresolved: [] },\n conflicts.length ? 'RECONCILED_WITH_CONFLICTS' : 'RECONCILED', conflicts.length ? 'supported' : 'proven',\n { question: asText(question), constraints: asText(constraints), authorityRule: 'freshest same-key assignment within supplied evidence' });\n }\n\n return resultEnvelope('RUMBO Deep Research Reconcile', IDS.drr,\n { facts: [], conclusion: null, reconciledConclusion: null, conflicts: [], superseded: [], unresolved: ['Evidence does not encode an authority-resolvable conclusion'] },\n 'FAIL_CLOSED_INSUFFICIENT_EVIDENCE', 'unknown', { question: asText(question), constraints: asText(constraints) });\n}\n\nexport function goalLoopController({ goal, currentState, blockers = '' }) {\n const currentJsonState = jsonInputState(currentState);\n if (currentJsonState.kind === 'INVALID') {\n return resultEnvelope('RUMBO Goal Loop Controller', IDS.glc,\n { nextAction: null, why: 'Invalid or ambiguous JSON current state', stopCondition: 'Provide strict duplicate-free JSON or unambiguous text state', openGates: ['INVALID_JSON_CURRENT_STATE'], jsonError: currentJsonState.error },\n 'FAIL_CLOSED_INVALID_JSON', 'unknown', { blockers: asText(blockers) || 'NONE', strictJsonValid: false });\n }\n const cj = tryJson(currentState);\n if (cj && typeof cj === 'object' && cj.nextAction) {\n return resultEnvelope('RUMBO Goal Loop Controller', IDS.glc,\n { nextAction: cj.nextAction, why: cj.why ?? 'Provided by current canonical state', stopCondition: cj.stopCondition ?? 'Observable goal state reached', openGates: cj.openGates ?? [] },\n 'NEXT_ACTION_SELECTED', 'proven', { blockers: asText(blockers) || 'NONE' });\n }\n const s = asText(currentState);\n const pending = s.match(/Acci[oó]n\\s+([A-Za-z0-9_-]+)\\s+es\\s+la\\s+[uú]nica\\s+acci[oó]n\\s+pendiente/i);\n const stop = s.match(/estado\\s+observable\\s+ser[aá]\\s+([A-Za-z0-9_-]+)/i) || s.match(/estado\\s+(?:final\\s+)?(?:ser[aá]|=)\\s*([A-Za-z0-9_-]+)/i);\n const blockerText = asText(blockers);\n const blockerLead = blockerText.split(/[.\\n;]/, 1)[0].trim();\n const hasBlocker = Boolean(blockerText) && !/^(ninguno|none|no blockers?|sin bloqueos?)$/i.test(blockerLead);\n if (hasBlocker) {\n return resultEnvelope('RUMBO Goal Loop Controller', IDS.glc,\n { nextAction: 'CLOSE_BLOCKER', why: asText(blockers), stopCondition: 'Blocker cleared with evidence', openGates: [blockerLead] },\n 'BLOCKED', 'supported', { duplicateAvoidance: true });\n }\n if (pending) {\n return resultEnvelope('RUMBO Goal Loop Controller', IDS.glc,\n { nextAction: pending[1], why: `${pending[1]} is the only evidenced pending action; completed actions must not be repeated`, stopCondition: stop ? `STATE=${stop[1]}` : 'Goal state observed', openGates: [] },\n 'NEXT_ACTION_SELECTED', 'proven', { duplicateAvoidance: true });\n }\n\n const assignments = extractAssignments(s);\n const negative = assignments.find(x => /^(?:FAIL(?:ED|_CLOSED)?|NO_GO|NOT_PROVEN|UNPROVEN|NOT_EXECUTED|PENDING|UNKNOWN)$/i.test(x.value));\n if (negative) {\n const gate = `${negative.key}=${negative.value}`;\n return resultEnvelope('RUMBO Goal Loop Controller', IDS.glc,\n { nextAction: `CLOSE_GATE: ${gate}`, why: `${gate} is the first evidence-supported open gate`, stopCondition: `Fresh evidence no longer reports ${gate}`, openGates: [gate] },\n 'NEXT_ACTION_SELECTED', 'supported', { duplicateAvoidance: true });\n }\n if (assignments.length) {\n return resultEnvelope('RUMBO Goal Loop Controller', IDS.glc,\n { nextAction: 'VERIFY_TARGET_STATE', why: 'Concrete current-state assignments are present and no open blocker is evidenced; verify the target state before promotion.', stopCondition: `Direct evidence proves the target goal: ${asText(goal)}`, openGates: [] },\n 'NEXT_ACTION_SELECTED', 'proven', { duplicateAvoidance: true });\n }\n return resultEnvelope('RUMBO Goal Loop Controller', IDS.glc,\n { nextAction: null, why: 'No unique evidence-grounded next action', stopCondition: 'Provide a unique pending gate or blocker', openGates: [] },\n 'FAIL_CLOSED_AMBIGUOUS_NEXT_ACTION', 'unknown', { goal: asText(goal) });\n}\n\nfunction resultCounts(text) {\n return [...asText(text).matchAll(/\\b(\\d+)\\s*\\/\\s*(\\d+)(?:\\s+|_)(PASS|FAIL)\\b/gi)]\n .map(m => ({ passed: Number(m[1]), total: Number(m[2]), status: m[3].toUpperCase() }));\n}\nfunction allObservedCountsPass(text) {\n const counts = resultCounts(text);\n return counts.length > 0 && counts.every(x => x.status === 'PASS' && x.passed === x.total);\n}\n\nexport function executeVerifyClose({ objective, implementation, verification = '' }) {\n const verificationJsonState = jsonInputState(verification);\n if (verificationJsonState.kind === 'INVALID') {\n return resultEnvelope('RUMBO Execute Verify Close', IDS.evc,\n { classification: 'INVALID_JSON_EVIDENCE', complete: false, evidenceSufficient: false, implementationComplete: false, promotionReady: false, externalGateStatusPresent: false, externalGatesClear: false, externalGateStatus: null, openGates: ['INVALID_JSON_EVIDENCE'], requiredEvidence: 'STRICT_DUPLICATE_FREE_JSON_OR_MACHINE_READABLE_TEXT', verificationState: asText(verification), missingEvidence: ['INVALID_JSON_EVIDENCE'], evidenceConflicts: [{ key: 'JSON', values: [verificationJsonState.error], reason: 'INVALID_OR_DUPLICATE_JSON' }], objective: asText(objective) },\n 'INVALID_JSON_EVIDENCE', 'unknown', { targetedTests: false, exactArtifact: false, regressions: false, externalGateStatusPresent: false, noExternalGates: false, conflictFree: false, strictJsonValid: false });\n }\n const v = asText(verification);\n const i = asText(implementation);\n const counts = resultCounts(v);\n\n const targetedState = evidenceCategoryState(v, 'targeted_or_smoke_test_pass');\n const exactState = evidenceCategoryState(v, 'exact_artifact_readback');\n const regressionState = evidenceCategoryState(v, 'regression_evidence');\n const externalCategory = externalGateCategory(v);\n\n const smokePass = /smoke(?:\\s+end-to-end)?\\s*[=:]?\\s*PASS/i.test(v);\n const explicitTestFailure = counts.some(x => x.status === 'FAIL' || x.passed !== x.total)\n || /\\b(?:test|tests|smoke|regresi[oó]n|regression|adversarial|static)[^.;\\n]{0,80}\\bFAIL\\b/i.test(v)\n || /\\b[A-Z][A-Z0-9_-]*\\s*=\\s*(?:FAIL|FAILED|NO_GO|NOT_PROVEN|UNPROVEN)\\b/i.test(v);\n const legacyPositive = !explicitTestFailure && (counts.length ? allObservedCountsPass(v) : smokePass || /(?:TESTS?|ADVERSARIAL|STATIC)\\s*=\\s*PASS/i.test(v));\n const positive = targetedState.present ? targetedState.positive : legacyPositive;\n\n const legacyExactArtifact = /(?:SOURCE_READBACK_SHA|DEPLOY_READBACK|READBACK_SHA)\\s*=\\s*PASS\\b/i.test(v)\n || (/SHA\\s*=\\s*[A-Za-z0-9_-]+/i.test(i + ' ' + v) && /readback[^.;\\n]{0,80}PASS/i.test(v));\n const exactArtifact = exactState.present ? exactState.positive : legacyExactArtifact;\n\n const legacyRegressions = counts.length >= 1 && allObservedCountsPass(v);\n const regressions = regressionState.present ? regressionState.positive : legacyRegressions;\n\n const legacyNoExternal = /(?:EXTERNAL_GATES?|REMOTE_EXTERNAL_GATES?)\\s*=\\s*NONE\\b/i.test(v)\n || /no\\s+hay\\s+gates?\\s+externos?\\s+pendientes?/i.test(v)\n || /external gates?\\s*[:=]\\s*none/i.test(v);\n const externalGateStatusPresent = externalCategory.present || legacyNoExternal;\n const noExternal = externalCategory.present ? externalCategory.clear : legacyNoExternal;\n const evidenceConflicts = [];\n if (targetedState.conflict || targetedState.invalid) evidenceConflicts.push({ key: 'targeted_or_smoke_test_pass', values: targetedState.values, reason: targetedState.conflict ? 'CONFLICTING_VALUES' : 'INVALID_VALUE' });\n if (exactState.conflict || exactState.invalid) evidenceConflicts.push({ key: 'exact_artifact_readback', values: exactState.values, reason: exactState.conflict ? 'CONFLICTING_VALUES' : 'INVALID_VALUE' });\n if (regressionState.conflict || regressionState.invalid) evidenceConflicts.push({ key: 'regression_evidence', values: regressionState.values, reason: regressionState.conflict ? 'CONFLICTING_VALUES' : 'INVALID_VALUE' });\n if (externalCategory.conflict) evidenceConflicts.push({ key: 'external_gate_status', values: externalCategory.values, reason: 'CONFLICTING_VALUES' });\n\n const missing = [];\n if (!positive) missing.push('targeted_or_smoke_test_pass');\n if (!exactArtifact) missing.push('exact_artifact_readback');\n if (!regressions) missing.push('regression_evidence');\n if (!externalGateStatusPresent) missing.push('external_gate_status');\n\n const conflictFree = evidenceConflicts.length === 0;\n const evidenceSufficient = missing.length === 0 && conflictFree;\n const implementationComplete = positive && exactArtifact && regressions && conflictFree;\n const promotionReady = evidenceSufficient && noExternal;\n const complete = promotionReady;\n const classification = evidenceConflicts.length\n ? 'CONFLICTING_EVIDENCE'\n : promotionReady\n ? 'RESUELTO_CON_EVIDENCIA'\n : evidenceSufficient\n ? 'VERIFICADO_GATE_EXTERNO_ABIERTO'\n : 'EJECUCION_SIN_EVIDENCIA';\n const openGates = evidenceConflicts.length\n ? ['EVIDENCE_CONFLICT']\n : missing.length ? [...missing] : (noExternal ? [] : ['EXTERNAL_GATE_OPEN']);\n\n return resultEnvelope('RUMBO Execute Verify Close', IDS.evc,\n {\n classification,\n complete,\n evidenceSufficient,\n implementationComplete,\n promotionReady,\n externalGateStatusPresent,\n externalGatesClear: noExternal,\n externalGateStatus: externalCategory.value,\n openGates,\n requiredEvidence: missing.length ? missing.join('; ') : 'NONE',\n verificationState: v,\n missingEvidence: missing,\n evidenceConflicts,\n objective: asText(objective),\n },\n classification, evidenceSufficient ? 'proven' : 'unknown',\n { targetedTests: positive, exactArtifact, regressions, externalGateStatusPresent, noExternalGates: noExternal, conflictFree });\n}\n\nconst NEGATIVE_STATUS_TOKEN_RE = /\\b(?:FAIL(?:ED|_CLOSED)?|NO_GO|NOT_PROVEN|UNPROVEN|NOT_EXECUTED|NOT_DEPLOYED|NOT_AUTHORIZED|NOT_EVALUATED|BLOCKED|PENDING|UNKNOWN|DENIED|REJECTED|NO_AUTHORITY|FALSE)\\b/ig;\n\nfunction explicitNegativeStatusTokens(text) {\n return [...new Set([...asText(text).matchAll(NEGATIVE_STATUS_TOKEN_RE)].map(m => m[0].toUpperCase()))];\n}\n\nfunction normalizedAssignmentValue(v) {\n return cleanToken(v).toUpperCase();\n}\n\nfunction distinctAssignmentValues(assignments, key) {\n return [...new Set(assignments\n .filter(x => x.key.toUpperCase() === key.toUpperCase())\n .map(x => normalizedAssignmentValue(x.value)))];\n}\n\nfunction negativeEvidenceAssignments(assignments) {\n const negative = /^(?:FAIL|FAILED|FAIL_CLOSED|NO_GO|NOT_PROVEN|UNPROVEN|NOT_EXECUTED|NOT_DEPLOYED|BLOCKED|PENDING|UNKNOWN|DENIED|REJECTED|NO_AUTHORITY|FALSE)$/i;\n return assignments.filter(x => negative.test(normalizedAssignmentValue(x.value)));\n}\n\nfunction assignmentConflicts(assignments) {\n const grouped = new Map();\n for (const item of assignments) {\n const key = item.key.toUpperCase();\n if (!grouped.has(key)) grouped.set(key, { key: item.key, values: [] });\n grouped.get(key).values.push(item.value);\n }\n const conflicts = [];\n for (const entry of grouped.values()) {\n const distinct = [...new Set(entry.values.map(normalizedAssignmentValue))];\n if (distinct.length > 1) conflicts.push({ key: entry.key, values: entry.values, reason: 'CONFLICTING_VALUES' });\n }\n return conflicts;\n}\n\nfunction conditionalGateRequirements(gates) {\n const g = asText(gates);\n const marker = g.match(/PASS\\s+(?:only if|[uú]nicamente\\s+si)\\s+(.+)/i);\n return marker ? extractAssignments(marker[1]) : [];\n}\n\nconst POSITIVE_STATUS_VALUE_RE = /^(?:PASS|PROVEN|VERIFIED|OK|TRUE|EXECUTED|DEPLOYED|AUTHORIZED|COMPLETE|COMPLETED)$/i;\nconst NEGATED_POSITIVE_RE = /(?:\\b(?:NO|NOT|DID\\s+NOT|DIDN['’]?T|FAILED\\s+TO|FAILURE\\s+TO|UNABLE\\s+TO|WITHOUT)\\b[^.;\\n]{0,64}\\b(?:PASS|PROVEN|VERIFIED|OK|EXECUTED|DEPLOYED|AUTHORIZED|SUCCEED(?:ED)?)\\b|\\b(?:PASS|PROVEN|VERIFIED|OK|EXECUTED|DEPLOYED|AUTHORIZED|SUCCEED(?:ED)?)\\b[^.;\\n]{0,64}\\b(?:NOT\\s+ACHIEVED|NOT\\s+PROVEN|NOT\\s+VERIFIED|FAILED|FALSE)\\b)/i;\n\nfunction claimStructureAssessment(text) {\n const input = asText(text);\n const state = jsonInputState(input);\n if (!input) return { structured: false, kind: 'EMPTY', residual: input };\n if (state.kind === 'INVALID') return { structured: false, kind: 'INVALID_JSON', residual: input };\n if (state.kind === 'VALID') {\n const assignments = flattenJsonScalarAssignments(state.value);\n return { structured: assignments.length > 0, kind: 'JSON', residual: assignments.length > 0 ? '' : input };\n }\n\n const pure = /^\\s*[A-Za-zÁÉÍÓÚÜÑáéíóúüñ_][\\wÁÉÍÓÚÜÑáéíóúüñ-]*\\s*=\\s*[^\\s,;]+(?:\\s*[;\\n]\\s*[A-Za-zÁÉÍÓÚÜÑáéíóúüñ_][\\wÁÉÍÓÚÜÑáéíóúüñ-]*\\s*=\\s*[^\\s,;]+)*\\s*[.;]?\\s*$/;\n const structured = pure.test(input);\n return { structured, kind: structured ? 'ASSIGNMENT_LIST' : 'UNSTRUCTURED_TEXT', residual: structured ? '' : input };\n}\n\nfunction positiveEvidenceAssessment(text) {\n const input = asText(text);\n const assignments = semanticAssignments(input);\n const positiveAssignments = assignments.filter(x => POSITIVE_STATUS_VALUE_RE.test(normalizedAssignmentValue(x.value)));\n const counts = resultCounts(input);\n const failedCounts = counts.filter(x => x.total <= 0 || x.status !== 'PASS' || x.passed !== x.total);\n const positiveCounts = counts.filter(x => x.total > 0 && x.status === 'PASS' && x.passed === x.total);\n const http200 = /\\bhttp_status\\s*[:=]\\s*200\\b/i.test(input);\n const statusOk = /\\bstatus\\s*[:=]\\s*ok\\b/i.test(input);\n const negatedPositive = NEGATED_POSITIVE_RE.test(input);\n const signals = [\n ...positiveAssignments.map(x => `${x.key}=${x.value}`),\n ...positiveCounts.map(x => `${x.passed}/${x.total}_${x.status}`),\n ...(http200 ? ['http_status=200'] : []),\n ...(statusOk ? ['status=ok'] : []),\n ];\n return {\n positive: !negatedPositive && failedCounts.length === 0 && signals.length > 0,\n signals,\n negatedPositive,\n failedCounts,\n };\n}\n\nexport function auditFinalState({ objective, claims, evidence, gates = '' }) {\n const objectiveText = asText(objective);\n const claimsText = asText(claims);\n const evidenceText = asText(evidence);\n const gatesText = asText(gates);\n\n const claimsJsonState = jsonInputState(claims);\n const evidenceJsonState = jsonInputState(evidence);\n const gatesJsonState = jsonInputState(gates);\n const invalidJson = [\n ...(claimsJsonState.kind === 'INVALID' ? [{ field: 'claims', error: claimsJsonState.error }] : []),\n ...(evidenceJsonState.kind === 'INVALID' ? [{ field: 'evidence', error: evidenceJsonState.error }] : []),\n ...(gatesJsonState.kind === 'INVALID' ? [{ field: 'gates', error: gatesJsonState.error }] : []),\n ];\n if (invalidJson.length) {\n return resultEnvelope('RUMBO Audit Final State', IDS.afs,\n { verdict: 'NO_GO', classification: 'UNPROVEN', nextGate: 'INVALID_JSON_INPUT', openGates: ['INVALID_JSON_INPUT'], conflicts: invalidJson.map(x => ({ key: x.field.toUpperCase(), values: [x.error], reason: 'INVALID_OR_DUPLICATE_JSON' })), objective: objectiveText, gateRequirements: [], claimRequirements: [], negativeEvidence: [], negativeEvidenceTokens: [], negativeClaimTokens: [], requiredGateSatisfied: null, claimBindingComplete: false },\n 'UNPROVEN', 'unknown', { objectivePresent: Boolean(objectiveText), claimsPresent: Boolean(claimsText), evidencePresent: Boolean(evidenceText), gatesPresent: Boolean(gatesText), explicitGatePolicy: false, gateRequiresSuccess: false, gatePolicyConflict: false, conditionalGateParsed: false, requiredGateSatisfied: null, claimBindingComplete: false, conflictFree: false, claimConflictFree: false, negativeEvidenceCount: 0, negativeClaimCount: 0, strictJsonValid: false, runtimeIdMustBeVerifiedExternally: IDS.afs });\n }\n\n const objectivePresent = Boolean(objectiveText);\n const claimsPresent = Boolean(claimsText);\n const evidencePresent = Boolean(evidenceText);\n const gatesPresent = Boolean(gatesText);\n\n const evidenceAssignments = semanticAssignments(evidenceText);\n const claimAssignments = semanticAssignments(claimsText);\n const claimStructure = claimStructureAssessment(claimsText);\n const negativeAssignments = negativeEvidenceAssignments(evidenceAssignments);\n const negativeEvidenceTokens = explicitNegativeStatusTokens(evidenceText);\n const evidenceConflicts = assignmentConflicts(evidenceAssignments);\n const claimNegativeAssignments = negativeEvidenceAssignments(claimAssignments);\n const negativeClaimTokens = explicitNegativeStatusTokens(claimsText);\n const claimConflicts = assignmentConflicts(claimAssignments);\n const evidenceConflictKeys = new Set(evidenceConflicts.map(x => x.key.toUpperCase()));\n const claimConflictKeys = new Set(claimConflicts.map(x => x.key.toUpperCase()));\n\n const claimRequirements = claimAssignments.map(req => {\n const matching = evidenceAssignments.filter(x => x.key.toUpperCase() === req.key.toUpperCase());\n const exact = matching.find(x => normalizedAssignmentValue(x.value) === normalizedAssignmentValue(req.value)) ?? null;\n const evidenceConflict = evidenceConflictKeys.has(req.key.toUpperCase());\n const claimConflict = claimConflictKeys.has(req.key.toUpperCase());\n return {\n key: req.key,\n expected: req.value,\n actual: matching.map(x => x.value),\n match: !evidenceConflict && !claimConflict && Boolean(exact),\n };\n });\n const claimBindingComplete = claimStructure.structured && claimAssignments.length > 0 && claimRequirements.every(x => x.match);\n const claimStateFailure = claimNegativeAssignments.length > 0 || negativeClaimTokens.length > 0 || claimConflicts.length > 0;\n\n const positiveAssessment = positiveEvidenceAssessment(evidenceText);\n const evidenceFailure = negativeEvidenceTokens.length > 0\n || negativeAssignments.length > 0\n || evidenceConflicts.length > 0\n || positiveAssessment.negatedPositive\n || positiveAssessment.failedCounts.length > 0;\n\n const gateObservedFailure = /\\b(?:gate|state|verdict|status)\\s*[:=]\\s*(?:FAIL(?:ED|_CLOSED)?|NO_GO|NOT_PROVEN|UNPROVEN)\\b/i.test(gatesText);\n const successEvidence = positiveAssessment.positive;\n const gateRequiresSuccess = /PASS\\s+(?:[uú]nicamente\\s+si|only if)/i.test(gatesText);\n const noneMarkerPresent = /(?:OPEN_GATES?|GATES?)\\s*=\\s*NONE\\b/i.test(gatesText);\n const explicitNoGates = /^(?:NONE|NINGUNO|SIN BLOQUEOS?|(?:OPEN_GATES?|GATES?)\\s*=\\s*NONE)\\.?$/i.test(gatesText.trim());\n const gatePolicyConflict = gateRequiresSuccess && noneMarkerPresent;\n const gatesNone = !gateRequiresSuccess && explicitNoGates;\n const explicitGatePolicy = gatesPresent && (gateRequiresSuccess || explicitNoGates);\n const requirements = conditionalGateRequirements(gatesText);\n const conflictKeys = new Set(evidenceConflicts.map(x => x.key.toUpperCase()));\n const gateRequirementStates = requirements.map(req => {\n const matching = evidenceAssignments.filter(x => x.key.toUpperCase() === req.key.toUpperCase());\n const actual = matching.at(-1) ?? null;\n const conflict = conflictKeys.has(req.key.toUpperCase());\n const match = !conflict && Boolean(actual) && normalizedAssignmentValue(actual.value) === normalizedAssignmentValue(req.value);\n return { key: req.key, expected: req.value, actual: actual?.value ?? null, match };\n });\n const requiredGateSatisfied = gateRequiresSuccess\n && requirements.length > 0\n && gateRequirementStates.every(x => x.match);\n\n const contextComplete = objectivePresent && claimsPresent && evidencePresent && gatesPresent;\n const verdict = contextComplete\n && explicitGatePolicy\n && claimStructure.structured\n && claimBindingComplete\n && !claimStateFailure\n && !evidenceFailure\n && !gateObservedFailure\n && !gatePolicyConflict\n && successEvidence\n && (gatesNone || requiredGateSatisfied)\n ? 'GO' : 'NO_GO';\n\n let nextGate = 'EVIDENCE_REQUIRED';\n if (!objectivePresent) nextGate = 'OBJECTIVE_REQUIRED';\n else if (!claimsPresent) nextGate = 'CLAIMS_REQUIRED';\n else if (!evidencePresent) nextGate = 'EVIDENCE_REQUIRED';\n else if (!gatesPresent || !explicitGatePolicy) nextGate = 'GATE_POLICY_REQUIRED';\n else if (!claimStructure.structured) nextGate = 'UNSTRUCTURED_CLAIM';\n else if (claimConflicts.length > 0) nextGate = 'CLAIM_CONFLICT';\n else if (claimNegativeAssignments.length > 0 || negativeClaimTokens.length > 0) nextGate = 'CLAIM_STATE_NOT_POSITIVE';\n else if (!claimBindingComplete) nextGate = 'CLAIM_EVIDENCE_BINDING_REQUIRED';\n else if (gatePolicyConflict) nextGate = 'GATE_POLICY_CONFLICT';\n else if (gateRequiresSuccess && !requiredGateSatisfied) nextGate = 'REQUIRED_GATE_NOT_PROVEN';\n else if (verdict === 'GO') nextGate = 'NONE';\n\n const classification = verdict === 'GO' ? 'PROVEN' : 'UNPROVEN';\n const conflicts = [\n ...evidenceConflicts,\n ...claimConflicts.map(x => ({ ...x, reason: 'CLAIM_CONFLICTING_VALUES' })),\n ...(gatePolicyConflict ? [{ key: 'GATE_POLICY', values: [gatesText], reason: 'NONE_AND_CONDITIONAL_REQUIREMENT' }] : []),\n ];\n\n return resultEnvelope('RUMBO Audit Final State', IDS.afs,\n {\n verdict,\n classification,\n nextGate,\n openGates: verdict === 'GO' ? [] : [nextGate],\n conflicts,\n objective: objectiveText,\n gateRequirements: gateRequirementStates,\n claimStructure,\n claimRequirements,\n claimBindingComplete,\n negativeClaims: claimNegativeAssignments.map(x => ({ key: x.key, value: x.value })),\n negativeClaimTokens,\n negativeEvidence: negativeAssignments.map(x => ({ key: x.key, value: x.value })),\n negativeEvidenceTokens,\n positiveEvidenceSignals: positiveAssessment.signals,\n failedResultCounts: positiveAssessment.failedCounts,\n negatedPositiveEvidence: positiveAssessment.negatedPositive,\n requiredGateSatisfied: gateRequiresSuccess ? requiredGateSatisfied : null,\n },\n classification, verdict === 'GO' ? 'proven' : 'unknown',\n {\n objectivePresent,\n claimsPresent,\n evidencePresent,\n gatesPresent,\n explicitGatePolicy,\n gateRequiresSuccess,\n gatePolicyConflict,\n conditionalGateParsed: !gateRequiresSuccess || requirements.length > 0,\n requiredGateSatisfied: gateRequiresSuccess ? requiredGateSatisfied : null,\n claimStructureKind: claimStructure.kind,\n structuredClaimsOnly: claimStructure.structured,\n claimBindingComplete,\n conflictFree: evidenceConflicts.length === 0,\n claimConflictFree: claimConflicts.length === 0,\n negativeEvidenceCount: Math.max(negativeAssignments.length, negativeEvidenceTokens.length) + positiveAssessment.failedCounts.length + (positiveAssessment.negatedPositive ? 1 : 0),\n negativeClaimCount: Math.max(claimNegativeAssignments.length, negativeClaimTokens.length),\n positiveEvidenceCount: positiveAssessment.signals.length,\n structuredPositiveEvidence: successEvidence,\n strictJsonValid: true,\n runtimeIdMustBeVerifiedExternally: IDS.afs,\n });\n}\n\nexport const semanticHandlers = Object.freeze({\n 'rumbo-canonical-state-recovery': canonicalStateRecovery,\n 'rumbo-deep-research-reconcile': deepResearchReconcile,\n 'rumbo-goal-loop-controller': goalLoopController,\n 'rumbo-execute-verify-close': executeVerifyClose,\n 'rumbo-audit-final-state': auditFinalState,\n});\n\nexport function mcpToolResult(output) {\n return {\n content: [{ type: 'text', text: JSON.stringify(output, null, 2) }],\n structuredContent: output,\n };\n}\n","label",[143],"f80d941a-bc93-4a4a-aa87-8ab1f6426931","rumbo-drift-sentinel.ts","01a099e2-134e-77a7-a880-20231c382c41",["D",1789288125263],"interval","import { blob } from \"https://esm.town/v/std/blob/main.ts\";\n\nconst EXPECTED = {\n runtime: \"V1.3.12\",\n main: {\n bytes: 3290,\n sha256: \"42bc0aaa25dddc5b11144d08f26228c1b39b15b36ee18e0718fcc32ae5f9406b\",\n },\n core: {\n bytes: 41135,\n sha256: \"5b143075ff78cd2cb4b55af85e4e540a0cc16262e79ebfcab97152e19df36f47\",\n },\n} as const;\n\nconst enc = new TextEncoder();\nconst hex = (buf: ArrayBuffer) =>\n Array.from(new Uint8Array(buf))\n .map((b) => b.toString(16).padStart(2, \"0\"))\n .join(\"\");\n\nasync function fingerprint(path: string) {\n const res = await fetch(new URL(path, import.meta.url), { cache: \"no-store\" });\n const text = await res.text();\n const bytes = enc.encode(text);\n return {\n status: res.status,\n bytes: bytes.length,\n sha256: hex(await crypto.subtle.digest(\"SHA-256\", bytes)),\n };\n}\n\nexport default async function (interval?: Interval) {\n const checkedAt = new Date().toISOString();\n const [main, core] = await Promise.all([\n fingerprint(\"./main.ts\"),\n fingerprint(\"./rumbo-semantic-core.mjs\"),\n ]);\n\n const checks = {\n mainStatus: main.status === 200,\n coreStatus: core.status === 200,\n mainBytes: main.bytes === EXPECTED.main.bytes,\n coreBytes: core.bytes === EXPECTED.core.bytes,\n mainSha256: main.sha256 === EXPECTED.main.sha256,\n coreSha256: core.sha256 === EXPECTED.core.sha256,\n };\n\n const failures = Object.entries(checks)\n .filter(([, ok]) => !ok)\n .map(([name]) => name);\n const verdict = failures.length === 0 ? \"PASS\" : \"DRIFT\";\n\n const receipt = {\n schema: \"rumbo-agent-reliability-drift-sentinel/v1\",\n checked_at_utc: checkedAt,\n previous_successful_run_at: interval?.lastRunAt?.toISOString() ?? null,\n runtime_expected: EXPECTED.runtime,\n verdict,\n failures,\n expected: EXPECTED,\n observed: { main, core },\n checks,\n };\n\n await blob.setJSON(\"rumbo/drift/latest.json\", receipt);\n await blob.setJSON(`rumbo/drift/history/${checkedAt}.json`, receipt);\n\n console.log(JSON.stringify(receipt));\n\n if (verdict !== \"PASS\") {\n throw new Error(`RUMBO_DRIFT_DETECTED:${failures.join(\",\")}`);\n }\n\n return receipt;\n}\n",[153],"0f6f541d-9be6-4e9a-ad1e-093c50128d50","main.ts","01a099e2-134e-77a7-a880-1e108f792b07",["D",1789288125263],"http","import { createMcpHandler, McpServer } from \"npm:@modelcontextprotocol/server\";\nimport * as z from \"npm:zod/v4\";\nimport {\n canonicalStateRecovery,\n deepResearchReconcile,\n goalLoopController,\n executeVerifyClose,\n auditFinalState,\n mcpToolResult,\n} from \"./rumbo-semantic-core.mjs\";\n\nfunction buildServer() {\n const server = new McpServer({ name: \"RUMBO Agent Reliability\", version: \"1.3.12\" });\n\n server.registerTool(\"rumbo-canonical-state-recovery\", {\n description: \"Recover one evidence-grounded canonical state before continuing long-running work. Read-only and fail-closed.\",\n inputSchema: z.object({ goal: z.string(), evidence: z.string(), priorState: z.string().optional() }),\n annotations: { readOnlyHint: true, destructiveHint: false, openWorldHint: false },\n }, async (args) => mcpToolResult(canonicalStateRecovery(args)));\n\n server.registerTool(\"rumbo-deep-research-reconcile\", {\n description: \"Turn supplied evidence into a reconciled answer while distinguishing facts, conflicts, and missing evidence.\",\n inputSchema: z.object({ question: z.string(), evidence: z.string(), constraints: z.string().optional() }),\n annotations: { readOnlyHint: true, destructiveHint: false, openWorldHint: false },\n }, async (args) => mcpToolResult(deepResearchReconcile(args)));\n\n server.registerTool(\"rumbo-goal-loop-controller\", {\n description: \"Choose the highest-value next action toward a goal without loops, duplicate work, or unsupported promotion.\",\n inputSchema: z.object({ goal: z.string(), currentState: z.string(), blockers: z.string().optional() }),\n annotations: { readOnlyHint: true, destructiveHint: false, openWorldHint: false },\n }, async (args) => mcpToolResult(goalLoopController(args)));\n\n server.registerTool(\"rumbo-execute-verify-close\", {\n description: \"Audit an implementation and define the exact evidence required before declaring it complete.\",\n inputSchema: z.object({ objective: z.string(), implementation: z.string(), verification: z.string().optional() }),\n annotations: { readOnlyHint: true, destructiveHint: false, openWorldHint: false },\n }, async (args) => mcpToolResult(executeVerifyClose(args)));\n\n server.registerTool(\"rumbo-audit-final-state\", {\n description: \"Perform a final fail-closed audit of claims, evidence, gates, and promotion readiness.\",\n inputSchema: z.object({ objective: z.string(), claims: z.string(), evidence: z.string(), gates: z.string().optional() }),\n annotations: { readOnlyHint: true, destructiveHint: false, openWorldHint: false },\n }, async (args) => mcpToolResult(auditFinalState(args)));\n\n return server;\n}\n\nconst handler = createMcpHandler(buildServer);\nexport default async function fetch(req: Request) {\n try {\n if (req.method === \"POST\") {\n const body = await req.clone().json();\n const method = typeof body?.method === \"string\" ? body.method : \"unknown\";\n const tool = typeof body?.params?.name === \"string\" ? body.params.name : null;\n console.log(JSON.stringify({ event: \"rumbo_mcp_request\", version: \"1.3.12+valtown-adapter.2\", method, tool }));\n }\n } catch {\n console.log(JSON.stringify({ event: \"rumbo_mcp_request\", version: \"1.3.12+valtown-adapter.2\", method: \"unparsed\", tool: null }));\n }\n return handler.fetch(req);\n}\n",[160],"9a4c80de-ef1a-449e-965e-01b381d05346","claim-evidence-admission.mjs","6ae1ab5e-af4d-11f1-b08a-1607ee4eb77e",5,["D",1789288246549],"import { sha256Canonical, stableJson, stableValue } from './vendor/rumbo-core/src/canonical.js';\nimport { validateEvidence, resolveSupersession } from './vendor/rumbo-core/src/evidence.js';\n\nexport const VERSION = 'RUMBO-CEA-V1.2.0-CANDIDATE';\nexport const CLAIM_BINDING_SCHEMA = 'rumbo.claim-binding/v1';\nexport const EVIDENCE_SCHEMA = 'rumbo-state-evidence/r1/evidence';\n\nfunction isPlainObject(v) {\n if (v === null || typeof v !== 'object' || Array.isArray(v)) return false;\n const proto = Object.getPrototypeOf(v);\n return proto === Object.prototype || proto === null;\n}\n\nfunction strictParseJson(text, field) {\n if (isPlainObject(text) || Array.isArray(text)) return text;\n if (typeof text !== 'string' || !text.trim()) throw new Error(`${field}:JSON_REQUIRED`);\n\n const s = text.trim();\n let i = 0;\n const fail = (code) => { throw new Error(`${field}:${code}`); };\n const ws = () => { while (i < s.length && /\\s/.test(s[i])) i += 1; };\n\n const readString = () => {\n if (s[i] !== '\"') fail('INVALID_JSON_STRING');\n const start = i;\n i += 1;\n while (i < s.length) {\n const ch = s[i];\n if (ch === '\"') {\n i += 1;\n try { return JSON.parse(s.slice(start, i)); } catch { fail('INVALID_JSON_STRING'); }\n }\n if (ch === '\\\\') {\n if (i + 1 >= s.length) fail('INVALID_JSON_ESCAPE');\n if (s[i + 1] === 'u') {\n if (!/^[0-9A-Fa-f]{4}$/.test(s.slice(i + 2, i + 6))) fail('INVALID_JSON_UNICODE_ESCAPE');\n i += 6;\n } else {\n if (!/[\"\\\\/bfnrt]/.test(s[i + 1])) fail('INVALID_JSON_ESCAPE');\n i += 2;\n }\n continue;\n }\n if (ch.charCodeAt(0) < 0x20) fail('INVALID_JSON_CONTROL_CHAR');\n i += 1;\n }\n fail('UNTERMINATED_JSON_STRING');\n };\n\n const readValue = () => {\n ws();\n if (i >= s.length) fail('UNEXPECTED_JSON_EOF');\n if (s[i] === '{') return readObject();\n if (s[i] === '[') return readArray();\n if (s[i] === '\"') { readString(); return; }\n const start = i;\n while (i < s.length && !/[\\s,\\]}]/.test(s[i])) i += 1;\n const token = s.slice(start, i);\n if (!token) fail('INVALID_JSON_VALUE');\n try { JSON.parse(token); } catch { fail('INVALID_JSON_VALUE'); }\n };\n\n const readObject = () => {\n i += 1;\n ws();\n const seen = new Set();\n if (s[i] === '}') { i += 1; return; }\n while (i < s.length) {\n ws();\n const key = readString();\n if (seen.has(key)) fail(`DUPLICATE_JSON_KEY:${key}`);\n seen.add(key);\n ws();\n if (s[i] !== ':') fail('INVALID_JSON_OBJECT_COLON');\n i += 1;\n readValue();\n ws();\n if (s[i] === '}') { i += 1; return; }\n if (s[i] !== ',') fail('INVALID_JSON_OBJECT_SEPARATOR');\n i += 1;\n }\n fail('UNTERMINATED_JSON_OBJECT');\n };\n\n const readArray = () => {\n i += 1;\n ws();\n if (s[i] === ']') { i += 1; return; }\n while (i < s.length) {\n readValue();\n ws();\n if (s[i] === ']') { i += 1; return; }\n if (s[i] !== ',') fail('INVALID_JSON_ARRAY_SEPARATOR');\n i += 1;\n }\n fail('UNTERMINATED_JSON_ARRAY');\n };\n\n readValue();\n ws();\n if (i !== s.length) fail('TRAILING_JSON_DATA');\n try { return JSON.parse(s); } catch { fail('INVALID_JSON'); }\n}\n\nfunction assertAllowedKeys(value, allowed, field) {\n if (!isPlainObject(value)) throw new Error(`${field}:OBJECT_REQUIRED`);\n for (const key of Object.keys(value)) {\n if (!allowed.includes(key)) throw new Error(`${field}:UNKNOWN_FIELD:${key}`);\n }\n}\n\nfunction canonicalId(v, field) {\n if (typeof v !== 'string' || !v.length) throw new Error(`${field}:ID_REQUIRED`);\n if (v !== v.trim()) throw new Error(`${field}:ID_NOT_CANONICAL`);\n return v;\n}\n\nfunction identity(v, field) {\n assertAllowedKeys(v, ['kind', 'id'], field);\n const kindRaw = typeof v.kind === 'string' ? v.kind : '';\n const idRaw = typeof v.id === 'string' ? v.id : '';\n if (!kindRaw || !idRaw) throw new Error(`${field}:KIND_AND_ID_REQUIRED`);\n if (kindRaw !== kindRaw.trim() || idRaw !== idRaw.trim()) throw new Error(`${field}:IDENTITY_NOT_CANONICAL`);\n const kind = kindRaw.toLowerCase();\n if (!/^[a-z][a-z0-9_.-]*$/.test(kind)) throw new Error(`${field}:INVALID_KIND`);\n return { kind, id: idRaw };\n}\n\nfunction sameIdentity(a, b) {\n return a?.kind === b?.kind && a?.id === b?.id;\n}\n\nfunction identityKey(v) {\n return `${v.kind}\\u0000${v.id}`;\n}\n\nfunction uniqueClaimIds(items) {\n const seen = new Set();\n for (const item of items) {\n if (!isPlainObject(item)) throw new Error('claims:ITEM_OBJECT_REQUIRED');\n const id = canonicalId(item.id, 'claim');\n if (seen.has(id)) throw new Error(`claims:DUPLICATE_ID:${id}`);\n seen.add(id);\n }\n}\n\nfunction uniqueEvidenceIds(records) {\n const seen = new Set();\n for (const record of records) {\n if (!isPlainObject(record)) throw new Error('evidence:RECORD_OBJECT_REQUIRED');\n const id = canonicalId(record.evidenceId, 'evidence');\n if (seen.has(id)) throw new Error(`evidence:DUPLICATE_ID:${id}`);\n seen.add(id);\n }\n}\n\nfunction normalizeClaim(cl, expectedIdentity = null) {\n assertAllowedKeys(cl, ['id', 'scope', 'predicate', 'value', 'subject', 'capability', 'evidence_refs'], 'claim');\n const id = canonicalId(cl.id, 'claim');\n const scope = typeof cl.scope === 'string' ? cl.scope.trim().toLowerCase() : '';\n if (!['target', 'capability'].includes(scope)) throw new Error(`claim.${id}:INVALID_SCOPE`);\n if (typeof cl.predicate !== 'string' || !cl.predicate.trim()) throw new Error(`claim.${id}:PREDICATE_REQUIRED`);\n if (cl.predicate !== cl.predicate.trim()) throw new Error(`claim.${id}:PREDICATE_NOT_CANONICAL`);\n if (!Object.hasOwn(cl, 'value')) throw new Error(`claim.${id}:VALUE_REQUIRED`);\n try { stableValue(cl.value); } catch { throw new Error(`claim.${id}:VALUE_NOT_CANONICALIZABLE`); }\n\n if (scope === 'target' && Object.hasOwn(cl, 'capability')) throw new Error(`claim.${id}:UNEXPECTED_CAPABILITY`);\n if (scope === 'capability' && Object.hasOwn(cl, 'subject')) throw new Error(`claim.${id}:UNEXPECTED_SUBJECT`);\n const claimIdentity = scope === 'target'\n ? identity(cl.subject, `claim.${id}.subject`)\n : identity(cl.capability, `claim.${id}.capability`);\n\n if (expectedIdentity && !sameIdentity(claimIdentity, expectedIdentity)) throw new Error(`claim.${id}:CLAIM_TARGET_MISMATCH`);\n if (!Array.isArray(cl.evidence_refs) || cl.evidence_refs.length === 0) throw new Error(`claim.${id}:EVIDENCE_REFS_REQUIRED`);\n const evidenceRefs = cl.evidence_refs.map((ref) => canonicalId(ref, `claim.${id}.evidence_ref`));\n if (new Set(evidenceRefs).size !== evidenceRefs.length) throw new Error(`claim.${id}:DUPLICATE_EVIDENCE_REF`);\n\n return {\n id,\n scope,\n predicate: cl.predicate,\n value: stableValue(cl.value),\n identity: claimIdentity,\n evidence_refs: evidenceRefs,\n };\n}\n\nexport function claimBindingDigest(claim) {\n const normalized = normalizeClaim(claim);\n return sha256Canonical({\n schema: CLAIM_BINDING_SCHEMA,\n claim_id: normalized.id,\n scope: normalized.scope,\n predicate: normalized.predicate,\n value: normalized.value,\n identity: normalized.identity,\n });\n}\n\nexport function claimToken(claim) {\n return `sha256:${claimBindingDigest(claim)}`;\n}\n\nfunction exclusionSet(negativeBindings) {\n const out = new Set();\n if (!negativeBindings) return out;\n const nb = strictParseJson(negativeBindings, 'negativeBindings');\n assertAllowedKeys(nb, ['subjects', 'capabilities'], 'negativeBindings');\n for (const key of ['subjects', 'capabilities']) {\n if (Object.hasOwn(nb, key) && !Array.isArray(nb[key])) throw new Error(`negativeBindings:${key.toUpperCase()}_ARRAY_REQUIRED`);\n }\n for (const raw of [...(nb.subjects ?? []), ...(nb.capabilities ?? [])]) out.add(identityKey(identity(raw, 'negativeBindings.identity')));\n return out;\n}\n\nfunction reject(id, reason, detail = {}) {\n return { id, admitted: false, reason, ...detail };\n}\n\nfunction nextGateFor(rejectedClaims, conflicts) {\n if (conflicts.length) return 'EVIDENCE_RECONCILIATION_REQUIRED';\n if (rejectedClaims.some((x) => x.reason === 'NEGATIVE_BINDING_EXCLUSION')) return 'RECOVERY_SCOPE_RECONCILIATION_REQUIRED';\n const nested = rejectedClaims.flatMap((x) => Array.isArray(x.problems) ? x.problems : []);\n if (nested.some((x) => x.reason === 'FOREIGN_EVIDENCE_IDENTITY' && x.scope === 'target')) return 'EXACT_SUBJECT_EVIDENCE_REQUIRED';\n if (nested.some((x) => x.reason === 'FOREIGN_EVIDENCE_IDENTITY' && x.scope === 'capability')) return 'EXACT_CAPABILITY_EVIDENCE_REQUIRED';\n if (nested.some((x) => x.reason === 'UNKNOWN_EVIDENCE_REF')) return 'EVIDENCE_REQUIRED';\n if (nested.some((x) => ['CLAIM_DIGEST_MISMATCH', 'EVIDENCE_VALUE_MISMATCH'].includes(x.reason))) return 'EXACT_CLAIM_ATTESTATION_REQUIRED';\n if (nested.some((x) => x.reason === 'SUPERSEDED_EVIDENCE_REF')) return 'FRESH_ACTIVE_EVIDENCE_REQUIRED';\n if (nested.some((x) => x.reason === 'EVIDENCE_CLASSIFICATION_NOT_ADMISSIBLE')) return 'ADMISSIBLE_EVIDENCE_CLASSIFICATION_REQUIRED';\n if (rejectedClaims.some((x) => x.reason === 'CLAIM_TARGET_MISMATCH')) return 'TARGET_IDENTITY_RECONCILIATION_REQUIRED';\n return 'INPUT_OR_CONTRACT_REPAIR_REQUIRED';\n}\n\nexport function evaluateClaimEvidenceAdmission({ target, claims, evidence, negativeBindings = '' }) {\n let parsedTarget;\n let parsedClaims;\n let parsedEvidence;\n let exclusions;\n try {\n parsedTarget = strictParseJson(target, 'target');\n parsedClaims = strictParseJson(claims, 'claims');\n parsedEvidence = strictParseJson(evidence, 'evidence');\n exclusions = exclusionSet(negativeBindings);\n } catch (err) {\n return finalize({\n verdict: 'NO_GO', classification: 'FAIL_CLOSED_INVALID_INPUT', nextGate: 'STRICT_INPUT_REQUIRED', target: null,\n admittedClaims: [], rejectedClaims: [reject(null, String(err?.message || err))], conflicts: [], evidenceRefs: [],\n });\n }\n\n try {\n assertAllowedKeys(parsedTarget, ['subject', 'capability'], 'target');\n assertAllowedKeys(parsedClaims, ['claims'], 'claims');\n assertAllowedKeys(parsedEvidence, ['records'], 'evidence');\n\n const subject = parsedTarget.subject ? identity(parsedTarget.subject, 'target.subject') : null;\n const capability = parsedTarget.capability ? identity(parsedTarget.capability, 'target.capability') : null;\n if (!subject && !capability) throw new Error('target:SUBJECT_OR_CAPABILITY_REQUIRED');\n\n const claimList = Array.isArray(parsedClaims.claims) ? parsedClaims.claims : null;\n const records = Array.isArray(parsedEvidence.records) ? parsedEvidence.records : null;\n if (!claimList?.length) throw new Error('claims:NONEMPTY_ARRAY_REQUIRED');\n if (!records?.length) throw new Error('evidence:NONEMPTY_RECORDS_REQUIRED');\n uniqueClaimIds(claimList);\n uniqueEvidenceIds(records);\n\n const normalizedClaims = new Map();\n const rejectedClaims = [];\n for (const raw of claimList) {\n try {\n const scope = typeof raw.scope === 'string' ? raw.scope.trim().toLowerCase() : '';\n const expected = scope === 'target' ? subject : scope === 'capability' ? capability : null;\n if (scope === 'target' && !subject) { rejectedClaims.push(reject(raw.id ?? null, 'TARGET_SUBJECT_NOT_DECLARED')); continue; }\n if (scope === 'capability' && !capability) { rejectedClaims.push(reject(raw.id ?? null, 'TARGET_CAPABILITY_NOT_DECLARED')); continue; }\n const cl = normalizeClaim(raw, expected);\n normalizedClaims.set(cl.id, { ...cl, claim_sha256: claimBindingDigest(raw), claim_token: claimToken(raw) });\n } catch (err) {\n const msg = String(err?.message || err);\n if (msg.includes('CLAIM_TARGET_MISMATCH')) rejectedClaims.push(reject(raw.id ?? null, 'CLAIM_TARGET_MISMATCH'));\n else rejectedClaims.push(reject(raw.id ?? null, 'INVALID_CLAIM_CONTRACT', { detail: msg }));\n }\n }\n\n const conflicts = [];\n const recordMap = new Map();\n for (const record of records) {\n const validation = validateEvidence(record);\n if (!validation.ok) {\n conflicts.push({ evidence_id: record?.evidenceId ?? null, reason: 'INVALID_R1_EVIDENCE', detail: validation.reason });\n continue;\n }\n recordMap.set(record.evidenceId, record);\n }\n\n let activeIds = new Set();\n let supersededIds = new Set();\n if (conflicts.length === 0) {\n const resolved = resolveSupersession(records);\n for (const link of resolved.invalidLinks) conflicts.push({ reason: 'INVALID_SUPERSESSION_LINK', ...link });\n for (const cycle of resolved.cycles) conflicts.push({ reason: 'SUPERSESSION_CYCLE', cycle });\n activeIds = new Set(resolved.active.map((record) => record.evidenceId));\n supersededIds = new Set(resolved.superseded.map((record) => record.evidenceId));\n }\n\n const admittedClaims = [];\n for (const [claimId, cl] of normalizedClaims.entries()) {\n if (exclusions.has(identityKey(cl.identity))) {\n rejectedClaims.push(reject(claimId, 'NEGATIVE_BINDING_EXCLUSION', { identity: cl.identity }));\n continue;\n }\n\n const problems = [];\n const used = [];\n for (const ref of cl.evidence_refs) {\n const record = recordMap.get(ref);\n if (!record) {\n problems.push({ ref, scope: cl.scope, reason: 'UNKNOWN_EVIDENCE_REF' });\n continue;\n }\n used.push(ref);\n\n if (supersededIds.has(ref)) problems.push({ ref, scope: cl.scope, reason: 'SUPERSEDED_EVIDENCE_REF' });\n else if (activeIds.size && !activeIds.has(ref)) problems.push({ ref, scope: cl.scope, reason: 'EVIDENCE_NOT_ACTIVE' });\n\n const actualIdentity = { kind: record.subject.kind, id: record.subject.id };\n if (!sameIdentity(actualIdentity, cl.identity)) {\n problems.push({ ref, scope: cl.scope, reason: 'FOREIGN_EVIDENCE_IDENTITY', expected: cl.identity, actual: actualIdentity });\n }\n\n if (record.subject.claim !== cl.claim_token) {\n problems.push({ ref, scope: cl.scope, reason: 'CLAIM_DIGEST_MISMATCH', expected_claim: cl.claim_token, actual_claim: record.subject.claim });\n }\n\n if (stableJson(record.after) !== stableJson(cl.value)) {\n problems.push({ ref, scope: cl.scope, reason: 'EVIDENCE_VALUE_MISMATCH', expected_value: cl.value, actual_value: record.after });\n }\n\n if (record.classification === 'NOT_PROVEN') {\n problems.push({ ref, scope: cl.scope, reason: 'EVIDENCE_CLASSIFICATION_NOT_ADMISSIBLE', classification: record.classification });\n }\n }\n\n if (problems.length) {\n rejectedClaims.push(reject(claimId, 'CLAIM_EVIDENCE_BINDING_FAILED', {\n scope: cl.scope, predicate: cl.predicate, value: cl.value, claim_sha256: cl.claim_sha256,\n evidenceRefs: used, problems,\n }));\n } else {\n admittedClaims.push({\n id: claimId, admitted: true, scope: cl.scope, predicate: cl.predicate, value: cl.value,\n identity: cl.identity, claim_sha256: cl.claim_sha256, claim_token: cl.claim_token,\n evidenceRefs: cl.evidence_refs,\n });\n }\n }\n\n if (conflicts.length) {\n for (const conflict of conflicts) rejectedClaims.push(reject(null, 'EVIDENCE_RECONCILIATION_CONFLICT', conflict));\n }\n\n const verdict = rejectedClaims.length === 0 && admittedClaims.length === claimList.length ? 'GO' : 'NO_GO';\n const nextGate = verdict === 'GO' ? 'NONE' : nextGateFor(rejectedClaims, conflicts);\n return finalize({\n verdict,\n classification: verdict === 'GO' ? 'R1_EVIDENCE_BOUND_TO_EXACT_SUBJECT_CLAIM_AND_VALUE' : 'UNPROVEN_OR_CROSS_BOUND_EVIDENCE',\n nextGate,\n target: { subject, capability },\n admittedClaims,\n rejectedClaims,\n conflicts,\n evidenceRefs: [...recordMap.keys()].sort(),\n canonicalEvidenceContract: EVIDENCE_SCHEMA,\n });\n } catch (err) {\n return finalize({\n verdict: 'NO_GO', classification: 'FAIL_CLOSED_INVALID_CONTRACT', nextGate: 'CONTRACT_REPAIR_REQUIRED', target: null,\n admittedClaims: [], rejectedClaims: [reject(null, String(err?.message || err))], conflicts: [], evidenceRefs: [],\n });\n }\n}\n\nfunction finalize(body) {\n const receiptBody = {\n schema: 'rumbo.claim-evidence-admission/v1.2',\n runtime_id: VERSION,\n authority: 'PRE_ANSWER_ADMISSION_ONLY_NO_PROMOTION_AUTHORITY',\n trust_boundary: 'R1_CANONICALITY_AND_HASH_INTEGRITY_DO_NOT_AUTHENTICATE_SOURCE_OR_AUTHORIZE_ACTION',\n required_outer_controls: ['SOURCE_TRUST_OR_AUTHENTICITY_WHEN_REQUIRED', 'LIVE_FRESHNESS_FOR_VOLATILE_STATE', 'SEPARATE_PROMOTION_AUTHORIZATION_FOR_CONSEQUENTIAL_ACTIONS'],\n ...body,\n };\n return {\n ...receiptBody,\n receipt_sha256: sha256Canonical(receiptBody),\n canonicalization: 'rumbo-core/stableJson',\n };\n}\n",[167],"cedcf130-c9ee-43d3-89f0-a3c84d241fe1","README.md","01a099e2-134e-77a7-a880-198fa8a08983",["D",1789288125263],"# MCP server\n\nA minimal MCP server with one example tool (that gets the current time in a\ngiven timezone).\n\nRemix this val to create your own MCP server on Val\nTown—[**_just like the MCP Core Maintainer team!_**](https://blog.val.town/mcp-core-maintainers)\n\nThis MCP server (<50 LOC in [`main.ts`](./main.ts)) is live on\n[mcp-time.val.run](https://mcp-time.val.run) if you'd like to test it\n([playground.ai.cloudflare.com](https://playground.ai.cloudflare.com) is a great\nplace to test an MCP server quickly). It's built on the v2 TypeScript SDK,\ncompatible with “MCP 2.0” (technically, the\n[2026-07-28 spec](https://modelcontextprotocol.io/specification/2026-07-28)). It\nalso supports the legacy spec, for MCP clients that haven't upgraded yet.",[174],"afafe617-e640-4826-a0a3-4f9fcd0f58f5","vendor","39e8cbae-af4d-11f1-9167-1607ee4eb77e",2,["D",1789288164382],"directory",[180],"pulls","branches",9,"forks","envVars",{"_194":61,"_196":61},"connections",{"_194":61,"_195":61},"count","unfilledCount","emptyCount",{"_42":179,"_31":180,"_29":181,"_145":-5,"_40":41,"_39":182,"_147":125,"_37":258,"_13":61,"_33":184,"_35":-5,"_23":24,"_150":-5,"_209":259,"_26":260},{"_42":253,"_31":233,"_29":248,"_145":181,"_40":41,"_39":254,"_147":125,"_37":255,"_13":61,"_33":184,"_35":-5,"_23":24,"_150":-5,"_209":256,"_26":257},{"_42":247,"_31":234,"_29":227,"_145":248,"_40":41,"_39":249,"_147":125,"_37":250,"_13":61,"_33":184,"_35":-5,"_23":24,"_150":-5,"_209":251,"_26":252},{"_42":239,"_31":240,"_29":241,"_145":227,"_40":41,"_39":242,"_147":125,"_37":243,"_13":61,"_33":34,"_35":244,"_23":24,"_150":-5,"_209":245,"_26":246},{"_42":235,"_31":32,"_29":30,"_145":227,"_40":41,"_39":14,"_147":125,"_37":236,"_13":61,"_33":34,"_35":36,"_23":24,"_150":-5,"_209":237,"_26":238},{"_42":224,"_31":225,"_29":226,"_145":227,"_40":41,"_39":228,"_147":125,"_37":229,"_13":61,"_33":34,"_35":230,"_23":24,"_150":-5,"_209":231,"_26":232},{"_42":173,"_31":174,"_29":175,"_145":-5,"_40":41,"_39":146,"_147":125,"_37":221,"_13":61,"_33":20,"_35":177,"_23":24,"_150":-5,"_209":222,"_26":223},{"_42":166,"_31":167,"_29":168,"_145":-5,"_40":41,"_39":169,"_147":24,"_37":218,"_13":61,"_33":20,"_35":171,"_23":24,"_150":-5,"_209":219,"_26":220},{"_42":159,"_31":160,"_29":161,"_145":-5,"_40":41,"_39":146,"_147":24,"_37":215,"_13":61,"_33":163,"_35":164,"_23":24,"_150":-5,"_209":216,"_26":217},{"_42":152,"_31":153,"_29":154,"_145":-5,"_40":41,"_39":146,"_147":24,"_37":212,"_13":61,"_33":156,"_35":157,"_23":24,"_150":-5,"_209":213,"_26":214},{"_42":142,"_31":143,"_29":144,"_145":-5,"_40":41,"_39":146,"_147":125,"_37":208,"_13":61,"_33":20,"_35":149,"_23":24,"_150":-5,"_209":210,"_26":211},["D",1789288125263],"children",[],[143],["D",1789288125263],[],[153],["D",1789288125263],[],[160],["D",1789288246549],[],[167],["D",1789288125263],[],[174],"becd0fa3-0d8b-437a-a692-2ac7b8858212","state.js","764b275e-af4d-11f1-9730-1607ee4eb77e","463c5be6-af4d-11f1-af77-1607ee4eb77e",7,["D",1789288265698],"export const STATE_EVIDENCE_VERSION = 'rumbo-state-evidence/r1';\n\nexport const DATA_CLASSIFICATIONS = Object.freeze([\n 'CANONICAL',\n 'OBSERVED',\n 'DERIVED',\n 'CACHED',\n 'NOT_PROVEN'\n]);\n\nexport const OPERATIONAL_STATES = Object.freeze([\n 'CURRENT',\n 'STALE',\n 'SUPERSEDED',\n 'CONFLICT',\n 'BLOCKED',\n 'NOT_PROVEN'\n]);\n\nexport function isDataClassification(value) {\n return DATA_CLASSIFICATIONS.includes(value);\n}\n\nexport function isOperationalState(value) {\n return OPERATIONAL_STATES.includes(value);\n}\n",[],[180,233,234,225],"rumbo-core","src","3efc3bd0-b05f-4324-9bef-621f23dae57e",["D",1789288273685],[],[180,233,234,32],"510d8bdb-6795-4850-b3ba-783f08753c6e","canonical.js","721a22ac-af4d-11f1-b93c-1607ee4eb77e",6,["D",1789288258672],"import { createHash } from 'node:crypto';\n\nfunction isPlainObject(value) {\n if (value === null || typeof value !== 'object') return false;\n const proto = Object.getPrototypeOf(value);\n return proto === Object.prototype || proto === null;\n}\n\nfunction assertDataProperty(object, key) {\n const descriptor = Object.getOwnPropertyDescriptor(object, key);\n if (!descriptor || !descriptor.enumerable || !Object.hasOwn(descriptor, 'value')) {\n throw new Error('INVALID_CANONICAL_VALUE: NON_DATA_PROPERTY');\n }\n return descriptor.value;\n}\n\nfunction stableArray(value, seen, visit) {\n if (Object.getOwnPropertySymbols(value).length) {\n throw new Error('INVALID_CANONICAL_VALUE: SYMBOL_KEY');\n }\n const names = Object.getOwnPropertyNames(value).filter((key) => key !== 'length');\n if (names.length !== value.length) throw new Error('INVALID_CANONICAL_VALUE: SPARSE_OR_EXTENDED_ARRAY');\n const output = [];\n for (let index = 0; index < value.length; index += 1) {\n if (!Object.hasOwn(value, index)) throw new Error('INVALID_CANONICAL_VALUE: SPARSE_ARRAY');\n output.push(visit(assertDataProperty(value, String(index)), seen));\n }\n return output;\n}\nfunction stableObject(value, seen, visit) {\n if (Object.getOwnPropertySymbols(value).length) {\n throw new Error('INVALID_CANONICAL_VALUE: SYMBOL_KEY');\n }\n const names = Object.getOwnPropertyNames(value).sort();\n const entries = names.map((key) => [key, visit(assertDataProperty(value, key), seen)]);\n return Object.fromEntries(entries);\n}\n\nfunction visitCanonical(value, seen) {\n if (value === null || typeof value === 'string' || typeof value === 'boolean') return value;\n if (typeof value === 'number') {\n if (!Number.isFinite(value)) throw new Error('INVALID_CANONICAL_VALUE: NON_FINITE_NUMBER');\n return value;\n }\n if (typeof value !== 'object') throw new Error(`INVALID_CANONICAL_VALUE: ${typeof value}`);\n if (seen.has(value)) throw new Error('CANONICAL_CYCLE');\n if (!Array.isArray(value) && !isPlainObject(value)) {\n throw new Error('INVALID_CANONICAL_VALUE: NON_PLAIN_OBJECT');\n }\n\n seen.add(value);\n try {\n return Array.isArray(value)\n ? stableArray(value, seen, visitCanonical)\n : stableObject(value, seen, visitCanonical);\n } finally {\n seen.delete(value);\n }\n}\nexport function stableValue(value) {\n return visitCanonical(value, new Set());\n}\n\nexport function stableJson(value) {\n return JSON.stringify(stableValue(value));\n}\n\nexport function sha256Canonical(value) {\n return createHash('sha256').update(stableJson(value), 'utf8').digest('hex');\n}\n",[],[180,233,234,240],"c047e875-9380-4a6e-ba13-17a1a58d32c8","3f1cbed2-af4d-11f1-b738-1607ee4eb77e",4,["D",1789288185079],[200,201,202],[180,233,234],"9e4c0e43-55ba-460c-904e-5310a696afbd",3,["D",1789288173056],[199],[180,233],["D",1789288164382],[198],[180],["D",1788582330391],{"_97":96,"_105":101},{"_97":96,"_98":99,"_100":101,"_18":54,"_102":103,"_104":24},{"_13":84,"_31":85,"_42":41,"_86":43,"_87":88,"_89":90,"_91":49,"_92":43,"_93":63,"_94":44,"_95":96},"ctx",{"_77":-5,"_127":298,"_299":300,"_301":302,"_303":24,"_304":24,"_305":24,"_306":125,"_307":-5,"_308":-7,"_309":-7,"_310":61,"_311":24,"_312":24,"_313":-5,"_314":24,"_315":24,"_316":24,"_317":-5},"ENV",{"_269":270,"_271":272,"_273":133,"_274":275,"_276":277,"_278":279,"_280":281,"_282":283,"_284":285,"_286":287,"_288":289,"_290":291,"_292":293,"_294":295,"_296":297},"REACT_APP_SENTRY_DSN","https://586020d86f05489db63a554690acf1e9@o1398359.ingest.sentry.io/6724499","RENDER_GIT_COMMIT","e87eb04e707cf64e7e532c8e8ec395f209e17f15","RENDER_GIT_BRANCH","POSTHOG_API_TOKEN","phc_nmoyuAkStEZuUp69rYlg1DZgrIt19Gx0Y4Hwf8tGgJP","POSTHOG_API_HOST","https://wart.val.town","WEB_API_TEMPLATE","https://{val}.web.val.run","APP_URL","https://www.val.town","MODULE_URL","https://esm.town","VANITY_WEB_API_TEMPLATE","https://{label}.val.run/","DENOLS_URL","https://lsp.val.town","NODE_ENV","production","CLOUDFLARE_TURNSTILE_SITE_KEY","0x4AAAAAACtaIZJH7s9rBMZ6","GOOGLE_CLIENT_ID","875989696899-jtsqkhlbdv5jrlsk8631p7ipjg3c6o62.apps.googleusercontent.com","STRIPE_PUBLISHABLE_KEY","pk_live_51N2zvWEpZTTLipIkVEwXrUG9AT0yeOOEYxszarZT1jpsZV58Ly6wlg3SltZpHN26CUalOBmUSY7adQRQiSkCZz2500qSXNkf8Y","API_URL","https://api.val.town",[],"preferencesKeybindings","default","preferencesTheme","light","preferencesFixers","preferencesCopilot","preferencesSelectionAi","preferencesCloseBrackets","preferencesChatPanelState","preferencesDprint","preferencesDenofmt","notificationCount","flagDenols","isBusinessOrgMember","businessOrgHandle","isBusinessOrgAdmin","isHigherTownieTier","navCollapsed","currentOrgId"]

Diff · v8