{ "name": "Contradiction_detection_agent", "nodes": [ { "parameters": {}, "id": "ecbef84d-9766-4795-a002-ae4a73dec262", "name": "Manual Trigger", "type": "n8n-nodes-base.manualTrigger", "typeVersion": 1, "position": [ -4672, 48 ] }, { "parameters": { "assignments": { "assignments": [ { "id": "s1", "name": "ticker", "value": "MSFT", "type": "string" }, { "id": "s2", "name": "company_name", "value": "Microsoft", "type": "string" }, { "id": "s3", "name": "lookback_days", "value": 365, "type": "number" }, { "id": "s4", "name": "fiscal_quarter", "value": "", "type": "string" }, { "id": "s5", "name": "active_patterns", "value": "=[1, 2, 3, 4, 5, 6]", "type": "array" }, { "id": "s6", "name": "resolve_id", "value": "=CONTRA-MSFT-{{ $now.toFormat('yyyyMMdd') }}-001", "type": "string" }, { "id": "s7", "name": "requested_at", "value": "={{ new Date().toISOString() }}", "type": "string" } ] }, "options": {} }, "id": "66032366-3c36-4dec-b28c-523200fe4def", "name": "Set Company Input", "type": "n8n-nodes-base.set", "typeVersion": 3.4, "position": [ -4448, 48 ] }, { "parameters": { "operation": "executeQuery", "query": "SELECT\n gs.id,\n gs.claim_id,\n gs.earnings_call_id,\n ec.ticker,\n ec.company_name,\n ec.call_date,\n ec.fiscal_quarter,\n gs.claim_type AS guidance_topic,\n gs.direction,\n gs.direction_confidence,\n gs.claim_text,\n gs.claim_summary,\n gs.transcript_citation,\n gs.change_description,\n gs.metric_name AS metric_type,\n gs.metric_value_current AS quantified_value,\n gs.metric_value_prior,\n gs.falsification_flag,\n gs.schema_valid\nFROM guidance_signals gs\nJOIN earnings_calls ec ON ec.id = gs.earnings_call_id\nWHERE ec.ticker = '{{ $('Set Company Input').first().json.ticker }}'\n AND gs.schema_valid = true\nORDER BY ec.call_date DESC, gs.direction_confidence DESC\nLIMIT 50;", "options": {} }, "id": "754ffa3c-de8a-4582-8f93-d80f44b93456", "name": "DB: Fetch Earnings Guidance Signals", "type": "n8n-nodes-base.postgres", "typeVersion": 2.5, "position": [ -3888, 80 ] }, { "parameters": { "operation": "executeQuery", "query": "SELECT\n ra.id,\n ra.claim_id,\n ra.earnings_call_id,\n ec.ticker,\n ec.call_date,\n ec.fiscal_quarter,\n ra.risk_category,\n ra.severity,\n ra.risk_text AS risk_description,\n ra.risk_summary,\n ra.uncertainty_language,\n ra.transcript_citation,\n ra.prior_disclosure_call,\n ra.schema_valid\nFROM risk_admissions ra\nJOIN earnings_calls ec ON ec.id = ra.earnings_call_id\nWHERE ec.ticker = '{{ $('Set Company Input').first().json.ticker }}'\n AND ra.schema_valid = true\nORDER BY ec.call_date DESC, ra.severity DESC\nLIMIT 30;", "options": {} }, "id": "abf60595-15d2-4c92-a7f1-8e6f13c69a4b", "name": "DB: Fetch Risk Admissions", "type": "n8n-nodes-base.postgres", "typeVersion": 2.5, "position": [ -3872, 544 ] }, { "parameters": { "operation": "executeQuery", "query": "SELECT\n qa.id,\n qa.earnings_call_id,\n ec.ticker,\n ec.call_date,\n ec.fiscal_quarter,\n qa.question_topic,\n qa.question_text,\n qa.response_type,\n qa.response_summary,\n qa.pressure_score,\n qa.asked_by_firms AS analyst_firm,\n qa.is_repeated_topic AS evasion_flag,\n qa.times_asked,\n qa.schema_valid\nFROM qa_pressure_map qa\nJOIN earnings_calls ec ON ec.id = qa.earnings_call_id\nWHERE ec.ticker = '{{ $('Set Company Input').first().json.ticker }}'\n AND qa.schema_valid = true\nORDER BY ec.call_date DESC, qa.pressure_score DESC\nLIMIT 30;", "options": {} }, "id": "5333b556-64a6-4914-8fdf-c46e6ca80820", "name": "DB: Fetch QA Pressure Map", "type": "n8n-nodes-base.postgres", "typeVersion": 2.5, "position": [ -3904, -128 ] }, { "parameters": { "operation": "executeQuery", "query": "SELECT\n ns.id,\n ns.claim_id,\n ns.ticker,\n ns.published_at,\n ns.source_name,\n ns.headline,\n ns.sentiment_label,\n ns.sentiment_score,\n ns.topic_tags,\n ns.schema_valid\nFROM news_signals ns\nWHERE ns.ticker = '{{ $('Set Company Input').first().json.ticker }}'\n AND ns.schema_valid = true\nORDER BY ns.published_at DESC\nLIMIT 50;", "options": {} }, "id": "94a1b236-3f97-4778-ad30-24b1e2ca70d8", "name": "DB: Fetch News Signals", "type": "n8n-nodes-base.postgres", "typeVersion": 2.5, "position": [ -3872, 400 ] }, { "parameters": { "operation": "executeQuery", "query": "SELECT\n tss.id,\n tss.claim_id,\n tss.ticker,\n tss.snapshot_date,\n tss.burst_detected,\n tss.burst_ratio,\n tss.velocity_anomaly,\n tss.velocity_z_score,\n tss.ai_focus_increasing,\n tss.rising_languages,\n tss.declining_languages,\n tss.emerging_languages,\n tss.stale_repo_count,\n tss.schema_valid\nFROM tech_stack_signals tss\nWHERE tss.ticker = '{{ $('Set Company Input').first().json.ticker }}'\n AND tss.schema_valid = true\nORDER BY tss.snapshot_date DESC\nLIMIT 10;", "options": {} }, "id": "e55c9080-c722-4f2f-88d0-1759b587e7a2", "name": "DB: Fetch Tech Stack Signals", "type": "n8n-nodes-base.postgres", "typeVersion": 2.5, "position": [ -3888, 224 ] }, { "parameters": { "jsCode": "const ctx = $('Set Company Input').first().json;\n\nfunction collectRows(nodeName) {\n try {\n return $(nodeName).all().map(i => i.json).filter(r => r && Object.keys(r).length > 0);\n } catch(e) { return []; }\n}\n\nconst guidance_signals = collectRows('DB: Fetch Earnings Guidance Signals').filter(r => r.claim_id);\nconst risk_admissions = collectRows('DB: Fetch Risk Admissions').filter(r => r.risk_category);\nconst qa_pressure = collectRows('DB: Fetch QA Pressure Map').filter(r => r.question_topic);\nconst news_signals = collectRows('DB: Fetch News Signals').filter(r => r.headline);\nconst tech_stack_signals = collectRows('DB: Fetch Tech Stack Signals').filter(r => r.ticker);\n\nconst news_count = news_signals.length;\nconst pos_news = news_signals.filter(n => n.sentiment_label === 'positive').length;\nconst neg_news = news_signals.filter(n => n.sentiment_label === 'negative').length;\nconst avg_news_score = news_count > 0\n ? news_signals.reduce((s, n) => s + (n.sentiment_score || 0), 0) / news_count\n : null;\nconst news_sentiment_direction = avg_news_score === null ? 'UNAVAILABLE'\n : avg_news_score > 0.2 ? 'POSITIVE'\n : avg_news_score < -0.2 ? 'NEGATIVE'\n : 'NEUTRAL';\n\nconst strengthened_guidance = guidance_signals.filter(g => g.direction === 'strengthened');\nconst weakened_guidance = guidance_signals.filter(g => g.direction === 'weakened');\nconst has_tech_stack = tech_stack_signals.length > 0;\nconst latest_tech = has_tech_stack ? tech_stack_signals[0] : null;\n\nreturn [{\n json: {\n ...ctx,\n signals: { guidance_signals, risk_admissions, qa_pressure, news_signals, tech_stack_signals },\n signal_counts: {\n guidance: guidance_signals.length,\n risks: risk_admissions.length,\n qa_items: qa_pressure.length,\n news: news_count,\n tech_stack: tech_stack_signals.length\n },\n summaries: {\n news_sentiment_direction,\n avg_news_score,\n pos_news,\n neg_news,\n strengthened_guidance_count: strengthened_guidance.length,\n weakened_guidance_count: weakened_guidance.length,\n unchanged_guidance_count: guidance_signals.filter(g => g.direction === 'unchanged').length,\n has_tech_stack,\n latest_tech\n }\n }\n}];" }, "id": "e1054950-8629-44ee-b485-946bf91db128", "name": "Aggregate All Signals", "type": "n8n-nodes-base.code", "typeVersion": 2, "position": [ -3440, 176 ] }, { "parameters": { "jsCode": "const ctx = $input.first().json;\nconst sigs = ctx.signals;\nconst summ = ctx.summaries;\nconst active = ctx.active_patterns || [1, 2, 3, 4, 5, 6];\nconst ticker = ctx.ticker;\nconst resolve_id = ctx.resolve_id;\n\nconst flags = [];\nconst skipped_patterns = [];\nconst pattern_results = {};\n\nlet flag_seq = 0;\nfunction makeFlag(pattern_id, pattern_name, severity, claim_a, claim_b, conflict_description, evidence_for_a, evidence_for_b) {\n flag_seq++;\n return {\n flag_id: `${resolve_id}-P${pattern_id}-F${String(flag_seq).padStart(3,'0')}`,\n resolve_id,\n ticker,\n pattern_id,\n pattern_name,\n severity,\n conflict_description,\n claim_a_id: claim_a?.claim_id || null,\n claim_a_source: claim_a?._source || null,\n claim_a_text: claim_a?.claim_text || claim_a?.headline || claim_a?.risk_description || null,\n claim_a_direction: claim_a?.direction || claim_a?.sentiment_label || null,\n claim_b_id: claim_b?.claim_id || null,\n claim_b_source: claim_b?._source || null,\n claim_b_text: claim_b?.claim_text || claim_b?.headline || null,\n claim_b_direction: claim_b?.direction || claim_b?.sentiment_label || null,\n evidence_for_a,\n evidence_for_b,\n resolution_status: 'UNRESOLVED',\n requires_human_review: true,\n schema_valid: true,\n flagged_at: new Date().toISOString()\n };\n}\n\n// PATTERN 1 — Sentiment vs Guidance Direction\nif (active.includes(1)) {\n const weakened = sigs.guidance_signals.filter(g => g.direction === 'weakened' && (g.direction_confidence || 0) >= 0.6);\n const news_pos = summ.news_sentiment_direction === 'POSITIVE';\n const news_neg = summ.news_sentiment_direction === 'NEGATIVE';\n const p1_flags = [];\n if (weakened.length > 0 && news_pos && (sigs.news_signals || []).length > 0) {\n for (const g of weakened.slice(0, 3)) {\n const best_pos = (sigs.news_signals || []).find(n => n.sentiment_label === 'positive');\n if (best_pos) {\n p1_flags.push(makeFlag(1, 'Sentiment vs Guidance Direction', 'HIGH',\n { ...g, _source: 'earnings_agent' },\n { ...best_pos, _source: 'news_agent' },\n `Guidance for \"${g.guidance_topic || g.claim_type || 'topic'}\" marked WEAKENED but news sentiment is POSITIVE`,\n `Earnings: direction=${g.direction}, confidence=${g.direction_confidence}`,\n `News: ${summ.pos_news} positive articles, avg score ${(summ.avg_news_score||0).toFixed(2)}`\n ));\n }\n }\n }\n if (summ.strengthened_guidance_count > 0 && news_neg && (sigs.news_signals||[]).length > 0) {\n const best_str = sigs.guidance_signals.find(g => g.direction === 'strengthened');\n const worst_neg = (sigs.news_signals||[]).find(n => n.sentiment_label === 'negative');\n if (best_str && worst_neg) {\n p1_flags.push(makeFlag(1, 'Sentiment vs Guidance Direction', 'HIGH',\n { ...best_str, _source: 'earnings_agent' },\n { ...worst_neg, _source: 'news_agent' },\n `Guidance STRENGTHENED on call but news sentiment is NEGATIVE`,\n `Earnings: ${summ.strengthened_guidance_count} strengthened signals`,\n `News: ${summ.neg_news} negative articles, avg score ${(summ.avg_news_score||0).toFixed(2)}`\n ));\n }\n }\n pattern_results[1] = { active: true, flags_raised: p1_flags.length };\n flags.push(...p1_flags);\n} else { skipped_patterns.push({ pattern_id: 1, reason: 'excluded_by_caller' }); }\n\n// PATTERN 2 — Risk Admission vs News Coverage\nif (active.includes(2)) {\n const high_risks = sigs.risk_admissions.filter(r => r.severity === 'high' || r.severity === 'critical');\n const p2_flags = [];\n for (const risk of high_risks.slice(0, 3)) {\n const matching_news = (sigs.news_signals||[]).filter(n =>\n (n.topic_tags||[]).some(t => risk.risk_category && t.toLowerCase().includes(risk.risk_category.toLowerCase()))\n );\n const positive_matching = matching_news.filter(n => n.sentiment_label === 'positive');\n if ((sigs.news_signals||[]).length > 0 && matching_news.length === 0) {\n p2_flags.push(makeFlag(2, 'Risk Admission vs News Coverage Gap', 'MEDIUM',\n { ...risk, _source: 'earnings_agent' }, null,\n `High-severity risk \"${risk.risk_category}\" admitted on call but absent from news coverage`,\n `Earnings: severity=${risk.severity}`,\n `News: ${(sigs.news_signals||[]).length} articles scanned, zero match topic`\n ));\n } else if (positive_matching.length > 0) {\n p2_flags.push(makeFlag(2, 'Risk Admission vs Coverage Tone', 'HIGH',\n { ...risk, _source: 'earnings_agent' },\n { ...positive_matching[0], _source: 'news_agent' },\n `High-severity risk \"${risk.risk_category}\" admitted on call; matching news has POSITIVE sentiment`,\n `Earnings: severity=${risk.severity}`,\n `News: sentiment=positive, headline=\"${(positive_matching[0].headline||'').slice(0,80)}\"`\n ));\n }\n }\n pattern_results[2] = { active: true, flags_raised: p2_flags.length };\n flags.push(...p2_flags);\n} else { skipped_patterns.push({ pattern_id: 2, reason: 'excluded_by_caller' }); }\n\n// PATTERN 3 — QA Evasion vs Analyst Coverage\nif (active.includes(3)) {\n const evaded = sigs.qa_pressure.filter(q => q.evasion_flag === true || q.evasion_flag === 'true' || q.pressure_score >= 7);\n const analyst_news = (sigs.news_signals||[]).filter(n =>\n (n.topic_tags||[]).some(t => ['analyst','rating','upgrade','target','price'].includes(t.toLowerCase()))\n );\n const pos_analyst = analyst_news.filter(n => n.sentiment_label === 'positive');\n const p3_flags = [];\n if (evaded.length >= 2 && pos_analyst.length > 0) {\n const top_evade = evaded.sort((a,b) => b.pressure_score - a.pressure_score)[0];\n const top_anews = pos_analyst[0];\n p3_flags.push(makeFlag(3, 'QA Evasion vs Analyst Confidence', 'MEDIUM',\n { ...top_evade, _source: 'earnings_agent', claim_id: `QA-${top_evade.earnings_call_id}-${top_evade.question_topic}`, claim_text: `Topic: ${top_evade.question_topic}, pressure_score: ${top_evade.pressure_score}` },\n { ...top_anews, _source: 'news_agent' },\n `Management evaded ${evaded.length} high-pressure Q&A topics but analyst coverage is positive`,\n `Q&A: ${evaded.length} evaded topics, top=\"${top_evade.question_topic}\"`,\n `Analyst news: sentiment=positive`\n ));\n }\n pattern_results[3] = { active: true, flags_raised: p3_flags.length };\n flags.push(...p3_flags);\n} else { skipped_patterns.push({ pattern_id: 3, reason: 'excluded_by_caller' }); }\n\n// PATTERN 4 — Tech Stack Decline vs Positive Guidance\nif (active.includes(4)) {\n if (!summ.has_tech_stack) {\n skipped_patterns.push({ pattern_id: 4, reason: 'tech_stack_data_unavailable' });\n pattern_results[4] = { active: false, flags_raised: 0, data_available: false };\n } else {\n const tech = summ.latest_tech;\n const declining = Array.isArray(tech.declining_languages) ? tech.declining_languages : [];\n const stale = tech.stale_repo_count || 0;\n const has_decline = declining.length >= 2 || stale >= 5;\n const pos_guidance = sigs.guidance_signals.filter(g => g.direction === 'strengthened' && (g.direction_confidence||0) >= 0.6);\n const p4_flags = [];\n if (has_decline && pos_guidance.length > 0) {\n const top_g = pos_guidance[0];\n p4_flags.push(makeFlag(4, 'Tech Stack Decline vs Positive Guidance', 'HIGH',\n { ...top_g, _source: 'earnings_agent' },\n { claim_id: `TS-${tech.ticker}-${tech.snapshot_date}`, _source: 'tech_stack_agent', claim_text: `Declining: ${declining.slice(0,3).join(', ')}; stale repos: ${stale}`, direction: 'declining' },\n `Management projects strengthened guidance but engineering footprint shows decline`,\n `Earnings: ${pos_guidance.length} strengthened signals, top=${top_g.guidance_topic}`,\n `Tech stack: ${declining.length} declining languages, ${stale} stale repos`\n ));\n }\n pattern_results[4] = { active: true, flags_raised: p4_flags.length, data_available: true };\n flags.push(...p4_flags);\n }\n} else { skipped_patterns.push({ pattern_id: 4, reason: 'excluded_by_caller' }); }\n\n// PATTERN 5 — Engineering Burst vs No Management Disclosure\nif (active.includes(5)) {\n if (!summ.has_tech_stack) {\n skipped_patterns.push({ pattern_id: 5, reason: 'tech_stack_data_unavailable' });\n pattern_results[5] = { active: false, flags_raised: 0, data_available: false };\n } else {\n const tech = summ.latest_tech;\n const has_burst = tech.burst_detected === true || tech.burst_detected === 'true';\n const has_velocity = tech.velocity_anomaly === true || tech.velocity_anomaly === 'true' || (tech.velocity_z_score || 0) >= 2.0;\n const emerging = Array.isArray(tech.emerging_languages) ? tech.emerging_languages : [];\n const pivot_signals = sigs.guidance_signals.filter(g =>\n g.direction === 'strengthened' ||\n (g.guidance_topic && ['strategy','pivot','expansion','new','direction'].some(kw =>\n (g.guidance_topic || '').toLowerCase().includes(kw)\n ))\n );\n const p5_flags = [];\n if ((has_burst || has_velocity) && pivot_signals.length === 0) {\n p5_flags.push(makeFlag(5, 'Engineering Burst vs Undisclosed Pivot', 'HIGH',\n null,\n { claim_id: `TS-${tech.ticker}-${tech.snapshot_date}`, _source: 'tech_stack_agent', claim_text: `Burst ratio=${tech.burst_ratio}, velocity z-score=${tech.velocity_z_score}, emerging: ${emerging.slice(0,3).join(', ')}`, direction: 'burst' },\n `Significant engineering burst detected but no corresponding management disclosure on earnings call`,\n `No earnings guidance matches the engineering direction change`,\n `Tech stack: burst=${has_burst}, velocity_anomaly=${has_velocity}, emerging=${emerging.join(', ')||'none'}`\n ));\n }\n pattern_results[5] = { active: true, flags_raised: p5_flags.length, data_available: true };\n flags.push(...p5_flags);\n }\n} else { skipped_patterns.push({ pattern_id: 5, reason: 'excluded_by_caller' }); }\n\n// PATTERN 6 — Guidance Optimism vs Negative News Momentum\nif (active.includes(6)) {\n const str_guidance = sigs.guidance_signals.filter(g => g.direction === 'strengthened' && (g.direction_confidence||0) >= 0.7);\n const news_neg_momentum = summ.news_sentiment_direction === 'NEGATIVE' && summ.neg_news > summ.pos_news && (sigs.news_signals||[]).length >= 5;\n const p6_flags = [];\n if (str_guidance.length > 0 && news_neg_momentum) {\n const top_str = str_guidance.sort((a,b) => (b.direction_confidence||0) - (a.direction_confidence||0))[0];\n const worst_n = (sigs.news_signals||[]).filter(n => n.sentiment_label === 'negative')[0];\n p6_flags.push(makeFlag(6, 'Guidance Optimism vs Negative News Momentum', 'HIGH',\n { ...top_str, _source: 'earnings_agent' },\n { ...worst_n, _source: 'news_agent' },\n `Management projects high-confidence strengthened guidance against sustained negative news momentum`,\n `Earnings: ${str_guidance.length} strengthened signals, confidence=${top_str.direction_confidence}`,\n `News: ${summ.neg_news} negative vs ${summ.pos_news} positive, avg=${(summ.avg_news_score||0).toFixed(2)}`\n ));\n }\n pattern_results[6] = { active: true, flags_raised: p6_flags.length };\n flags.push(...p6_flags);\n} else { skipped_patterns.push({ pattern_id: 6, reason: 'excluded_by_caller' }); }\n\nconst has_contradictions = flags.length > 0;\nconst high_severity_count = flags.filter(f => f.severity === 'HIGH').length;\nconst medium_severity_count = flags.filter(f => f.severity === 'MEDIUM').length;\nconst overall_confidence_level =\n high_severity_count >= 2 ? 'CRITICALLY_COMPROMISED'\n : high_severity_count >= 1 ? 'CONFIDENCE_REDUCED'\n : medium_severity_count >= 2 ? 'REVIEW_WARRANTED'\n : !has_contradictions ? 'NO_CONTRADICTIONS_DETECTED'\n : 'LOW_CONCERN';\n\nreturn [{\n json: {\n ...ctx,\n contradiction_flags: flags,\n pattern_results,\n skipped_patterns,\n detection_summary: {\n total_flags: flags.length,\n high_severity: high_severity_count,\n medium_severity: medium_severity_count,\n low_severity: flags.filter(f => f.severity === 'LOW').length,\n patterns_active: Object.keys(pattern_results).filter(k => pattern_results[k].active).length,\n patterns_skipped: skipped_patterns.length,\n has_contradictions,\n overall_confidence_level,\n requires_human_review: has_contradictions\n }\n }\n}];" }, "id": "cab09bc8-9ae5-4388-93f9-a905486c0ddd", "name": "Run Pattern Detection Engine", "type": "n8n-nodes-base.code", "typeVersion": 2, "position": [ -3264, 160 ] }, { "parameters": { "jsCode": "const ctx = $input.first().json;\nconst flags = ctx.contradiction_flags || [];\nconst summ = ctx.summaries;\nconst sigCounts = ctx.signal_counts;\nconst detSumm = ctx.detection_summary;\n\nif (!detSumm.has_contradictions) {\n return [{ json: { ...ctx, llm_needed: false } }];\n}\n\nconst flagSummary = flags.map((f, i) =>\n `[Flag ${i+1}] Pattern ${f.pattern_id} — ${f.pattern_name} | Severity: ${f.severity}\\n` +\n `Conflict: ${f.conflict_description}\\n` +\n `Evidence A (${f.claim_a_source||'N/A'}): ${(f.evidence_for_a||'').slice(0,200)}\\n` +\n `Evidence B (${f.claim_b_source||'N/A'}): ${(f.evidence_for_b||'').slice(0,200)}`\n).join('\\n\\n');\n\nconst system_prompt = `You are a senior investment intelligence analyst for the Mycroft platform. You receive structured contradiction flags produced by automated detection logic. Your job is to: 1. Assess the plausibility of each contradiction. 2. Assign an investment relevance score (1-10) to each flag. 3. Identify which contradictions are self-resolving vs genuinely irreconcilable. 4. Produce a concise analyst memo (3-5 sentences) per flag. 5. Produce a single overall portfolio-level assessment paragraph. Rules: Never recommend buy/sell/hold. Surface conflict, do not resolve it. Flag INSUFFICIENT_EVIDENCE when data is thin. Cite only evidence provided. Return ONLY valid JSON. No markdown, no preamble.`;\n\nconst user_prompt = `TICKER: ${ctx.ticker} | COMPANY: ${ctx.company_name} | RESOLVE_ID: ${ctx.resolve_id}\\n\\nSIGNAL INVENTORY:\\n- Guidance signals: ${sigCounts.guidance}\\n- Risk admissions: ${sigCounts.risks}\\n- QA pressure items: ${sigCounts.qa_items}\\n- News articles: ${sigCounts.news}\\n- Tech stack snapshots: ${sigCounts.tech_stack}\\n\\nNEWS SENTIMENT: ${summ.news_sentiment_direction}\\n\\nDETECTED CONTRADICTION FLAGS (${flags.length} total):\\n\\n${flagSummary}\\n\\nReturn JSON:\\n{\\n \"overall_assessment\": \"\",\\n \"overall_investment_relevance\": <1-10>,\\n \"flags\": [\\n {\\n \"flag_id\": \"\",\\n \"plausibility\": \"REAL_CONFLICT | TIMING_ARTEFACT | FRAMING_DIFFERENCE | INSUFFICIENT_EVIDENCE\",\\n \"investment_relevance_score\": <1-10>,\\n \"self_resolving\": ,\\n \"analyst_memo\": \"<3-5 sentences>\",\\n \"recommended_action\": \"ESCALATE_TO_HUMAN | MONITOR | DISMISS\"\\n }\\n ]\\n}`;\n\nreturn [{\n json: {\n ...ctx,\n llm_needed: true,\n groq_payload: {\n model: 'llama-3.3-70b-versatile',\n max_tokens: 2000,\n temperature: 0.1,\n messages: [\n { role: 'system', content: system_prompt },\n { role: 'user', content: user_prompt }\n ]\n }\n }\n}];" }, "id": "a74eaeef-e39b-47a5-9468-7ccf43f3da0f", "name": "Build Groq Prompt", "type": "n8n-nodes-base.code", "typeVersion": 2, "position": [ -3056, 160 ] }, { "parameters": { "conditions": { "options": { "caseSensitive": true, "leftValue": "", "typeValidation": "strict", "version": 1 }, "conditions": [ { "id": "cond-llm", "leftValue": "={{ $json.llm_needed }}", "rightValue": true, "operator": { "type": "boolean", "operation": "equals" } } ], "combinator": "and" }, "options": {} }, "id": "cafae45c-034a-4c5b-aeb7-cb0ccc5e1093", "name": "LLM Needed?", "type": "n8n-nodes-base.if", "typeVersion": 2, "position": [ -2896, 160 ] }, { "parameters": { "method": "POST", "url": "https://api.groq.com/openai/v1/chat/completions", "authentication": "predefinedCredentialType", "nodeCredentialType": "groqApi", "sendBody": true, "bodyParameters": { "parameters": [ {} ] }, "options": { "response": { "response": { "neverError": true } } } }, "id": "d9d5878a-5b33-4994-9955-8f00069ff9e9", "name": "Groq: Analyse Contradictions", "type": "n8n-nodes-base.httpRequest", "typeVersion": 4.2, "position": [ -2592, 128 ] }, { "parameters": { "jsCode": "const groqResp = $input.first().json;\nconst ctx = $('Build Groq Prompt').first().json;\n\nconst rawText = groqResp.choices?.[0]?.message?.content || '{}';\nlet llm_analysis;\ntry {\n const clean = rawText.replace(/```json|```/g, '').trim();\n llm_analysis = JSON.parse(clean);\n} catch(e) {\n llm_analysis = { overall_assessment: 'LLM_PARSE_FAILED: ' + e.message, overall_investment_relevance: null, flags: [], parse_error: true };\n}\n\nconst enriched_flags = (ctx.contradiction_flags || []).map(flag => {\n const llm_flag = (llm_analysis.flags || []).find(lf => lf.flag_id === flag.flag_id) || {};\n return {\n ...flag,\n plausibility: llm_flag.plausibility || 'INSUFFICIENT_EVIDENCE',\n investment_relevance_score: llm_flag.investment_relevance_score || null,\n self_resolving: llm_flag.self_resolving ?? null,\n analyst_memo: llm_flag.analyst_memo || null,\n recommended_action: llm_flag.recommended_action || 'ESCALATE_TO_HUMAN'\n };\n});\n\nconst run_log = {\n resolve_id: ctx.resolve_id,\n run_type: 'contradiction_llm_analysis',\n model_used: 'llama-3.3-70b-versatile',\n prompt_tokens: groqResp.usage?.prompt_tokens || 0,\n completion_tokens: groqResp.usage?.completion_tokens || 0,\n success: !llm_analysis.parse_error\n};\n\nreturn [{\n json: {\n ...ctx,\n contradiction_flags: enriched_flags,\n llm_analysis: {\n overall_assessment: llm_analysis.overall_assessment,\n overall_investment_relevance: llm_analysis.overall_investment_relevance\n },\n run_log\n }\n}];" }, "id": "0b7be9a9-6b8d-417e-ae6b-a09dde49e7c3", "name": "Process Groq Response", "type": "n8n-nodes-base.code", "typeVersion": 2, "position": [ -2400, 128 ] }, { "parameters": { "jsCode": "const ctx = $input.first().json;\nreturn [{\n json: {\n ...ctx,\n llm_analysis: {\n overall_assessment: 'No contradictions detected. All signals are consistent within defined patterns.',\n overall_investment_relevance: 0\n },\n run_log: {\n resolve_id: ctx.resolve_id,\n run_type: 'contradiction_no_flags',\n model_used: 'none',\n prompt_tokens: 0,\n completion_tokens: 0,\n success: true\n }\n }\n}];" }, "id": "9aa0d618-18b2-4b72-bd93-38bceccb58d2", "name": "No-Flag Passthrough", "type": "n8n-nodes-base.code", "typeVersion": 2, "position": [ -2592, 320 ] }, { "parameters": { "operation": "executeQuery", "query": "INSERT INTO contradiction_reports (\n resolve_id, ticker, company_name, fiscal_quarter,\n lookback_days, total_flags, high_severity_count,\n medium_severity_count, overall_confidence_level,\n overall_assessment, overall_investment_relevance,\n patterns_active, patterns_skipped, signal_counts,\n requires_human_review, created_at\n) VALUES (\n '{{ $json.resolve_id }}',\n '{{ $json.ticker }}',\n '{{ $json.company_name }}',\n {{ $json.fiscal_quarter ? \"'\" + $json.fiscal_quarter + \"'\" : 'NULL' }},\n {{ $json.lookback_days }},\n {{ $json.detection_summary.total_flags }},\n {{ $json.detection_summary.high_severity }},\n {{ $json.detection_summary.medium_severity }},\n '{{ $json.detection_summary.overall_confidence_level }}',\n '{{ ($json.llm_analysis && $json.llm_analysis.overall_assessment) ? $json.llm_analysis.overall_assessment.replace(/'/g, \"''\").slice(0,1000) : \"\" }}',\n {{ ($json.llm_analysis && $json.llm_analysis.overall_investment_relevance != null && $json.llm_analysis.overall_investment_relevance !== undefined) ? $json.llm_analysis.overall_investment_relevance : 'NULL' }},\n {{ $json.detection_summary.patterns_active }},\n {{ $json.detection_summary.patterns_skipped }},\n '{{ JSON.stringify($json.signal_counts) }}'::jsonb,\n {{ $json.detection_summary.requires_human_review ? 'true' : 'false' }},\n NOW()\n)\nON CONFLICT (resolve_id) DO UPDATE SET\n total_flags = EXCLUDED.total_flags,\n overall_confidence_level = EXCLUDED.overall_confidence_level\nRETURNING id, resolve_id, overall_confidence_level;", "options": {} }, "id": "b5a94d34-7511-4ebe-93de-b4e8a620f5b1", "name": "DB: Insert Contradiction Report", "type": "n8n-nodes-base.postgres", "typeVersion": 2.5, "position": [ -2368, 272 ] }, { "parameters": { "jsCode": "// Fan out each flag for individual DB insertion\nlet flags = [];\ntry { flags = $('Process Groq Response').first().json.contradiction_flags || []; } catch(e) {}\nif (!flags.length) {\n try { flags = $('No-Flag Passthrough').first().json.contradiction_flags || []; } catch(e) {}\n}\nif (flags.length === 0) {\n return [{ json: { no_flags: true } }];\n}\nreturn flags.map(f => ({ json: f }));" }, "id": "8519f7cd-1a7a-4f0e-8bc5-618839ccdf5f", "name": "Fan Out Flags", "type": "n8n-nodes-base.code", "typeVersion": 2, "position": [ -2112, 256 ], "alwaysOutputData": true }, { "parameters": { "operation": "executeQuery", "query": "INSERT INTO contradiction_flags (\n flag_id, resolve_id, ticker, pattern_id, pattern_name,\n severity, conflict_description,\n claim_a_id, claim_a_source, claim_a_text, claim_a_direction,\n claim_b_id, claim_b_source, claim_b_text, claim_b_direction,\n evidence_for_a, evidence_for_b,\n plausibility, investment_relevance_score, self_resolving,\n analyst_memo, recommended_action,\n resolution_status, requires_human_review, flagged_at\n) VALUES (\n '{{ $json.flag_id }}',\n '{{ $json.resolve_id }}',\n '{{ $json.ticker }}',\n {{ $json.pattern_id }},\n '{{ $json.pattern_name }}',\n '{{ $json.severity }}',\n '{{ ($json.conflict_description||'').replace(/'/g,\"''\").slice(0,500) }}',\n {{ $json.claim_a_id ? \"'\" + $json.claim_a_id + \"'\" : 'NULL' }},\n {{ $json.claim_a_source ? \"'\" + $json.claim_a_source + \"'\" : 'NULL' }},\n {{ $json.claim_a_text ? \"'\" + ($json.claim_a_text||'').replace(/'/g,\"''\").slice(0,500) + \"'\" : 'NULL' }},\n {{ $json.claim_a_direction ? \"'\" + $json.claim_a_direction + \"'\" : 'NULL' }},\n {{ $json.claim_b_id ? \"'\" + $json.claim_b_id + \"'\" : 'NULL' }},\n {{ $json.claim_b_source ? \"'\" + $json.claim_b_source + \"'\" : 'NULL' }},\n {{ $json.claim_b_text ? \"'\" + ($json.claim_b_text||'').replace(/'/g,\"''\").slice(0,500) + \"'\" : 'NULL' }},\n {{ $json.claim_b_direction ? \"'\" + $json.claim_b_direction + \"'\" : 'NULL' }},\n '{{ ($json.evidence_for_a||'').replace(/'/g,\"''\").slice(0,600) }}',\n '{{ ($json.evidence_for_b||'').replace(/'/g,\"''\").slice(0,600) }}',\n '{{ $json.plausibility || 'INSUFFICIENT_EVIDENCE' }}',\n {{ $json.investment_relevance_score !== null && $json.investment_relevance_score !== undefined ? $json.investment_relevance_score : 'NULL' }},\n {{ $json.self_resolving !== null && $json.self_resolving !== undefined ? $json.self_resolving : 'NULL' }},\n {{ $json.analyst_memo ? \"'\" + ($json.analyst_memo||'').replace(/'/g,\"''\").slice(0,1000) + \"'\" : 'NULL' }},\n '{{ $json.recommended_action || 'ESCALATE_TO_HUMAN' }}',\n '{{ $json.resolution_status || 'UNRESOLVED' }}',\n {{ $json.requires_human_review ? 'true' : 'false' }},\n '{{ $json.flagged_at }}'\n)\nON CONFLICT (flag_id) DO NOTHING\nRETURNING id, flag_id, severity;", "options": {} }, "id": "86d3bcb6-a793-41fe-9645-2ac3b76a639e", "name": "DB: Insert Contradiction Flag", "type": "n8n-nodes-base.postgres", "typeVersion": 2.5, "position": [ -1888, 240 ] }, { "parameters": { "jsCode": "// Build clean final report visible in n8n output panel\nlet ctx;\ntry { ctx = $('Process Groq Response').first().json; } catch(e) {}\nif (!ctx) { ctx = $('No-Flag Passthrough').first().json; }\n\nconst detS = ctx.detection_summary;\nconst sigC = ctx.signal_counts;\nconst flags = ctx.contradiction_flags || [];\nconst llmA = ctx.llm_analysis || {};\n\nconst report = {\n resolve_id: ctx.resolve_id,\n ticker: ctx.ticker,\n company_name: ctx.company_name,\n generated_at: new Date().toISOString(),\n\n detection_summary: {\n total_flags: detS.total_flags,\n high_severity: detS.high_severity,\n medium_severity: detS.medium_severity,\n overall_confidence_level: detS.overall_confidence_level,\n patterns_active: detS.patterns_active,\n patterns_skipped: detS.patterns_skipped,\n requires_human_review: detS.requires_human_review\n },\n\n signal_inventory: {\n guidance_signals: sigC.guidance,\n risk_admissions: sigC.risks,\n qa_pressure_items: sigC.qa_items,\n news_articles: sigC.news,\n tech_stack_snapshots: sigC.tech_stack\n },\n\n overall_assessment: llmA.overall_assessment,\n overall_investment_relevance: llmA.overall_investment_relevance,\n\n contradiction_flags: flags.map(f => ({\n flag_id: f.flag_id,\n pattern_id: f.pattern_id,\n pattern_name: f.pattern_name,\n severity: f.severity,\n conflict_description: f.conflict_description,\n plausibility: f.plausibility,\n investment_relevance_score: f.investment_relevance_score,\n recommended_action: f.recommended_action,\n analyst_memo: f.analyst_memo,\n requires_human_review: f.requires_human_review\n })),\n\n deferred_patterns: (ctx.skipped_patterns || []).map(p => ({\n pattern_id: p.pattern_id,\n reason: p.reason\n }))\n};\n\nreturn [{ json: report }];" }, "id": "d34eff1d-fbf8-4a9b-9b11-defe19ef85e5", "name": "Build Final Report", "type": "n8n-nodes-base.code", "typeVersion": 2, "position": [ -1648, 240 ] }, { "parameters": { "operation": "executeQuery", "query": "CREATE TABLE IF NOT EXISTS contradiction_reports (\n id UUID DEFAULT gen_random_uuid() PRIMARY KEY,\n resolve_id TEXT UNIQUE NOT NULL,\n ticker TEXT,\n company_name TEXT,\n fiscal_quarter TEXT,\n lookback_days INT,\n total_flags INT DEFAULT 0,\n high_severity_count INT DEFAULT 0,\n medium_severity_count INT DEFAULT 0,\n overall_confidence_level TEXT,\n overall_assessment TEXT,\n overall_investment_relevance INT,\n patterns_active INT,\n patterns_skipped INT,\n signal_counts JSONB,\n requires_human_review BOOLEAN DEFAULT TRUE,\n created_at TIMESTAMPTZ DEFAULT NOW()\n);", "options": {} }, "type": "n8n-nodes-base.postgres", "typeVersion": 2.6, "position": [ -3888, 848 ], "id": "6dbdc1f0-eb42-4c68-9f4a-8978eb41678d", "name": "Execute a SQL query", "disabled": true }, { "parameters": { "operation": "executeQuery", "query": "CREATE TABLE IF NOT EXISTS contradiction_flags (\n id UUID DEFAULT gen_random_uuid() PRIMARY KEY,\n flag_id TEXT UNIQUE NOT NULL,\n resolve_id TEXT,\n ticker TEXT,\n pattern_id INT,\n pattern_name TEXT,\n severity TEXT,\n conflict_description TEXT,\n claim_a_id TEXT,\n claim_a_source TEXT,\n claim_a_text TEXT,\n claim_a_direction TEXT,\n claim_b_id TEXT,\n claim_b_source TEXT,\n claim_b_text TEXT,\n claim_b_direction TEXT,\n evidence_for_a TEXT,\n evidence_for_b TEXT,\n plausibility TEXT,\n investment_relevance_score INT,\n self_resolving BOOLEAN,\n analyst_memo TEXT,\n recommended_action TEXT,\n resolution_status TEXT DEFAULT 'UNRESOLVED',\n requires_human_review BOOLEAN DEFAULT TRUE,\n flagged_at TIMESTAMPTZ\n);", "options": {} }, "type": "n8n-nodes-base.postgres", "typeVersion": 2.6, "position": [ -3712, 880 ], "id": "9ca58f8f-035c-4ccc-aeee-f129204a3207", "name": "Execute a SQL query1", "disabled": true }, { "parameters": { "operation": "executeQuery", "query": "SELECT ticker, burst_detected, \n burst_ratio, velocity_anomaly,\n declining_languages\nFROM tech_stack_signals\nWHERE ticker = 'MSFT';", "options": {} }, "type": "n8n-nodes-base.postgres", "typeVersion": 2.6, "position": [ -3216, 880 ], "id": "a31e1d6f-f9b6-4291-adb5-b05edf50fab3", "name": "Execute a SQL query2", "disabled": true }, { "parameters": { "operation": "executeQuery", "query": "CREATE TABLE IF NOT EXISTS news_signals (\n id UUID DEFAULT gen_random_uuid() PRIMARY KEY,\n claim_id TEXT,\n ticker TEXT,\n published_at TIMESTAMPTZ,\n source_name TEXT,\n headline TEXT,\n sentiment_label TEXT,\n sentiment_score NUMERIC,\n topic_tags TEXT[],\n schema_valid BOOLEAN DEFAULT TRUE\n);\nALTER TABLE news_signals \nADD CONSTRAINT news_signals_claim_id_unique \nUNIQUE (claim_id);", "options": {} }, "type": "n8n-nodes-base.postgres", "typeVersion": 2.6, "position": [ -3504, 880 ], "id": "9ed47953-8b26-4575-a28c-277ce2583072", "name": "Execute a SQL query3", "disabled": true }, { "parameters": { "url": "https://www.alphavantage.co/query", "sendQuery": true, "queryParameters": { "parameters": [ { "name": "function", "value": "NEWS_SENTIMENT" }, { "name": "tickers ", "value": "={{ $('Set Company Input').first().json.ticker }}" }, { "name": "limit", "value": "50" }, { "name": "sort", "value": "LATEST" }, { "name": "apikey", "value": "REPLACE_WITH_APIKEY" } ] }, "options": {} }, "type": "n8n-nodes-base.httpRequest", "typeVersion": 4.2, "position": [ -4400, 272 ], "id": "17604ba3-410c-4e21-81a9-7b77a38e41d0", "name": "HTTP Request" }, { "parameters": { "jsCode": "const response = $input.first().json;\nconst ticker = $('Set Company Input').first().json.ticker;\n\nconst feed = response.feed || [];\n\nif (feed.length === 0) {\n return [{ json: { news_empty: true, ticker } }];\n}\n\nconst processed = feed.map((article, idx) => {\n // Find sentiment for our specific ticker\n const tickerSentiment = (article.ticker_sentiment || []).find(\n t => t.ticker === ticker\n );\n\n const sentimentScore = tickerSentiment\n ? parseFloat(tickerSentiment.ticker_sentiment_score || 0)\n : parseFloat(article.overall_sentiment_score || 0);\n\n const sentimentLabel = sentimentScore > 0.15 ? 'positive'\n : sentimentScore < -0.15 ? 'negative'\n : 'neutral';\n\n // Extract topics\n const topic_tags = (article.topics || []).map(t => t.topic.toLowerCase());\n\n return {\n claim_id: `NEWS-${ticker}-${idx + 1}`,\n ticker,\n published_at: article.time_published\n ? new Date(\n article.time_published.replace(\n /(\\d{4})(\\d{2})(\\d{2})T(\\d{2})(\\d{2})(\\d{2})/,\n '$1-$2-$3T$4:$5:$6'\n )\n ).toISOString()\n : new Date().toISOString(),\n source_name: article.source || 'unknown',\n headline: (article.title || '').replace(/'/g, \"''\").slice(0, 300),\n sentiment_label: sentimentLabel,\n sentiment_score: sentimentScore,\n topic_tags,\n schema_valid: true\n };\n});\n\nreturn processed.map(n => ({ json: n }));" }, "type": "n8n-nodes-base.code", "typeVersion": 2, "position": [ -4240, 272 ], "id": "691e8e71-2a21-49ed-bd7d-4a3d598731bd", "name": "Process News Response" }, { "parameters": { "operation": "executeQuery", "query": "INSERT INTO news_signals(\n claim_id, ticker, published_at, source_name,\n headline, sentiment_label, sentiment_score,\n topic_tags, schema_valid\n)\nVALUES (\n '{{ $json.claim_id }}',\n '{{ $json.ticker }}',\n '{{ $json.published_at }}'::TIMESTAMPTZ,\n '{{ ($json.source_name || \"\").replace(/'/g, \" \").replace(/'/g, \" \").replace(/'/g, \" \") }}',\n '{{ ($json.headline || \"\").replace(/'/g, \" \").replace(/'/g, \" \").replace(/'/g, \" \") }}',\n '{{ $json.sentiment_label }}',\n {{ $json.sentiment_score }},\n ARRAY[{{ ($json.topic_tags || []).map(t => \"'\" + t.replace(/'/g, \" \") + \"'\").join(',') || \"'general'\" }}],\n true\n)\nON CONFLICT (claim_id) DO UPDATE SET\n sentiment_label = EXCLUDED.sentiment_label,\n sentiment_score = EXCLUDED.sentiment_score;", "options": {} }, "type": "n8n-nodes-base.postgres", "typeVersion": 2.6, "position": [ -4064, 288 ], "id": "3ca2bddd-f0e8-4695-bea3-c56892addc38", "name": "DB: Save News Signals" }, { "parameters": { "numberInputs": 5 }, "type": "n8n-nodes-base.merge", "typeVersion": 3.2, "position": [ -3584, 144 ], "id": "dc819982-436b-4db5-941a-262891e19e77", "name": "Merge" }, { "id": "24668fbf-09a9-538e-91d7-b5d62ae4cb47", "name": "Publication · Overview and setup", "type": "n8n-nodes-base.stickyNote", "typeVersion": 1, "position": [ -5472, -128 ], "parameters": { "content": "# Contradiction_detection_agent\n\n## Purpose and structure\nThis folder contains the **Contradiction_detection_agent** workflow. Its executable nodes use `code`, `httpRequest`, `if`, `manualTrigger`, `merge`, `postgres`, `set`. The table and connection map below are derived from the JSON, not an inferred guarantee of behavior.\n\n## Setup\nImport inactive. Configure your own credentials and replace REPLACE_WITH_* values. Review external destinations and Code nodes before running.\n\n## Entry points\nManual Trigger\n\n## Verification\nSanitized source candidate; not runtime-tested. This overview is generated from node metadata. Read the folder README for the connection map, credential types, and license.\n\n## Attribution\nNik Bear Brown; rights confirmed 2026-10-19\nMIT; preserve source notices.", "width": 700, "height": 800, "color": 1 } } ], "connections": { "Manual Trigger": { "main": [ [ { "node": "Set Company Input", "type": "main", "index": 0 } ] ] }, "Set Company Input": { "main": [ [ { "node": "HTTP Request", "type": "main", "index": 0 }, { "node": "DB: Fetch Earnings Guidance Signals", "type": "main", "index": 0 }, { "node": "DB: Fetch Tech Stack Signals", "type": "main", "index": 0 }, { "node": "DB: Fetch Risk Admissions", "type": "main", "index": 0 }, { "node": "DB: Fetch QA Pressure Map", "type": "main", "index": 0 } ] ] }, "DB: Fetch Earnings Guidance Signals": { "main": [ [ { "node": "Merge", "type": "main", "index": 1 } ] ] }, "DB: Fetch Risk Admissions": { "main": [ [ { "node": "Merge", "type": "main", "index": 4 } ] ] }, "DB: Fetch QA Pressure Map": { "main": [ [ { "node": "Merge", "type": "main", "index": 0 } ] ] }, "DB: Fetch News Signals": { "main": [ [ { "node": "Merge", "type": "main", "index": 3 } ] ] }, "DB: Fetch Tech Stack Signals": { "main": [ [ { "node": "Merge", "type": "main", "index": 2 } ] ] }, "Aggregate All Signals": { "main": [ [ { "node": "Run Pattern Detection Engine", "type": "main", "index": 0 } ] ] }, "Run Pattern Detection Engine": { "main": [ [ { "node": "Build Groq Prompt", "type": "main", "index": 0 } ] ] }, "Build Groq Prompt": { "main": [ [ { "node": "LLM Needed?", "type": "main", "index": 0 } ] ] }, "LLM Needed?": { "main": [ [ { "node": "Groq: Analyse Contradictions", "type": "main", "index": 0 } ], [ { "node": "No-Flag Passthrough", "type": "main", "index": 0 } ] ] }, "Groq: Analyse Contradictions": { "main": [ [ { "node": "Process Groq Response", "type": "main", "index": 0 } ] ] }, "Process Groq Response": { "main": [ [ { "node": "DB: Insert Contradiction Report", "type": "main", "index": 0 } ] ] }, "No-Flag Passthrough": { "main": [ [ { "node": "DB: Insert Contradiction Report", "type": "main", "index": 0 } ] ] }, "DB: Insert Contradiction Report": { "main": [ [ { "node": "Fan Out Flags", "type": "main", "index": 0 } ] ] }, "Fan Out Flags": { "main": [ [ { "node": "DB: Insert Contradiction Flag", "type": "main", "index": 0 } ] ] }, "DB: Insert Contradiction Flag": { "main": [ [ { "node": "Build Final Report", "type": "main", "index": 0 } ] ] }, "HTTP Request": { "main": [ [ { "node": "Process News Response", "type": "main", "index": 0 } ] ] }, "Process News Response": { "main": [ [ { "node": "DB: Save News Signals", "type": "main", "index": 0 } ] ] }, "DB: Save News Signals": { "main": [ [ { "node": "DB: Fetch News Signals", "type": "main", "index": 0 } ] ] }, "Merge": { "main": [ [ { "node": "Aggregate All Signals", "type": "main", "index": 0 } ] ] } }, "active": false, "settings": { "executionOrder": "v1" } }