"raw_score": 0.92,
      "weighted_contribution": 0.23,
      "evidence_id": "ticket_2026-03-01_001",
      "event_fingerprint": "sha1-of-title-and-tickers"
    },
    {
      "skill": "theme_detector",
      "signal_ref": "theme_ai_infra",
      "raw_score": 0.85,
      "weighted_contribution": 0.13
    }
  ],
  "tickers": ["NVDA", "AMD", "AVGO"],
  "direction": "LONG",
  "time_horizon": "3-6 months",
  "confidence_breakdown": {
    "multi_skill_agreement": 0.30,
    "signal_strength": 0.35,
    "recency": 0.22
  }
}

], "contradictions": [ { "contradiction_id": "contra_001", "description": "Conflicting sector view on Energy", "skill_a": { "skill": "sector_analyst", "signal": "Energy sector bearish rotation", "direction": "SHORT" }, "skill_b": { "skill": "institutional_flow_tracker", "signal": "Heavy institutional buying in XLE", "direction": "LONG" }, "resolution_hint": "Check timeframe mismatch (short-term vs long-term)" } ], "deduplication_log": [ { "merged_into": "edge_candidate_agent:ticket_001", "duplicates_removed": ["edge_candidate_agent:ticket_001@reports/ticket_001.json"], "similarity_score": 1.0, "reason": "same evidence re-ingested", "selection_rule": "newest timestamp wins; ties, missing, or unparsable timestamps keep the highest raw_score, then a stable key", "retained_copy": { "signal_ref": "ticket_001", "source_file": "reports/rerun.json", "raw_score": 0.5, "timestamp": "2026-10-10T00:00:00Z" }, "discarded_copies": [ { "signal_ref": "ticket_001", "source_file": "reports/ticket_001.json", "raw_score": 0.9, "timestamp": "2026-10-09T00:00:00Z" } ] }, { "merged_into": "edge_candidate_agent:ticket_001", "duplicates_removed": ["theme_detector:ai_compute"], "similarity_score": 0.9 } ], "invalid_signals": [ { "skill": "theme_detector", "signal_ref": "broken_theme", "title": "Theme with a corrupted score", "reason": "non_finite", "declared_scale": null, "raw_value_repr": "nan", "source_file": "reports/theme_detector_2026-03-01.json" } ], "scale_assumptions_report": [ { "skill": "theme_detector", "signal_ref": "ai_compute", "value": 7.2, "normalized_value": 0.072, "assumed_scale": "0-100", "source_file": "reports/theme_detector.json" } ] }

### Producer contract (optional keys, schema 1.1)

Each producer item (ticket, concept, theme, sector row, flow, hint) MAY carry:

- `evidence_id` — a stable identifier for the underlying evidence. Without this key a
  deterministic fallback is derived from skill, signal reference, title, tickers, and direction
  (timestamps are excluded so refreshed timestamps cannot forge a new identity). Re-ingesting
  the same evidence (same skill + `evidence_id`) never adds a bonus: all copies collapse into
  one. When the copies differ, the **newest observation wins**: the copy with the latest
  timestamp, compared as parsed timezone-aware datetimes (date-only and naive values are read
  as UTC), becomes the retained score. Equal, missing, or unparsable timestamps fall back to
  the highest `raw_score`, then a stable key (`signal_ref`, `source_file`), so ingestion order
  never changes the result. Only ISO-8601 strings (or `YYYY-MM-DD`) are ordered; epoch numbers and compact forms such as `20261010` count as missing timestamps. Remaining full ties are broken by title, direction, tickers, and a canonical dump of the signal. A copy that carries a score older than the retained one is
  discarded, so a newer, lower score replaces an older, higher one. The discarded copies'
  scores and timestamps are listed in `deduplication_log[].discarded_copies`. Identical
  copies leave the score unchanged.
- `score_scale` — `"0-1"` or `"0-100"`, declaring the raw score's scale. Declared scales are
  enforced strictly: NaN/infinity/negative/out-of-range values are rejected into
  `invalid_signals`, never converted into maximum confidence. A present-but-unparsable
  declaration (`""`, `"foo"`) is also rejected (`missing_scale`). If the key is absent, values
  in (1, 100] are divided by 100 and flagged `score_scale_assumed: true` plus a
  `scale_assumptions_report` entry (`value` is the original input, `normalized_value` the
  0-1 score used); values > 100 are rejected.

### Agreement rule (Issue #510)

Agreement bonuses fire on `agreement_tier = min(distinct evidence_ids, distinct
event_fingerprints, distinct skills)` among a merged signal's contributors: tier 2 earns the
two-skills bonus, tier 3+ the three-plus bonus. Re-ingested duplicates earn nothing; the same
event restated by another skill (identical normalized title + tickers) counts once. Known
limitation: the event fingerprint is vocabulary-based, so paraphrased restatements with disjoint
word sets still earn the agreement bonus — treat cross-skill paraphrases with independent
verification. Merged duplicates no longer earn a per-copy bonus (`merge_bonus_per_duplicate`
is deprecated and ignored).

### Markdown Report

The markdown report provides a human-readable dashboard:

```markdown
# Edge Signal Aggregator Dashboard
**Generated:** 2026-03-02 07:00 UTC

## Summary
- Total Input Signals: 42
- Unique After Dedup: 28
- Contradictions: 3
- High Conviction (>=0.5): 12

## Top 10 Edge Ideas by Conviction

### 1. AI Infrastructure Capex Acceleration (Score: 0.87)
- **Tickers:** NVDA, AMD, AVGO
- **Direction:** LONG | **Horizon:** 3-6 months
- **Contributing Skills:**
  - edge-candidate-agent: 0.92 (ticket_2026-03-01_001)
  - theme-detector: 0.85 (theme_ai_infra)
- **Confidence Breakdown:** Agreement 0.30 | Strength 0.35 | Recency 0.22

...

## Contradictions Requiring Review

### Energy Sector Conflict
- **sector-analyst:** Bearish rotation (SHORT)
- **institutional-flow-tracker:** Heavy buying XLE (LONG)
- **Hint:** Check timeframe mismatch

## Deduplication Summary
- 14 signals merged into 8 unique themes
- Average similarity of merged signals: 0.89

Reports are saved to reports/ with filenames:

• edge_signal_aggregator_YYYY-MM-DD_HHMMSS.json
• edge_signal_aggregator_YYYY-MM-DD_HHMMSS.md

Resources

• scripts/aggregate_signals.py -- Main aggregation script with CLI interface
• references/signal-weighting-framework.md -- Rationale for default weights and scoring methodology
• assets/default_weights.yaml -- Default skill weights configuration

Key Principles

1. Provenance Tracking -- Every aggregated signal links back to its source skill and original reference
2. Contradiction Transparency -- Conflicting signals are flagged, not hidden, to enable informed decisions
3. Configurable Weights -- Default weights reflect typical reliability but can be customized per user
4. Deduplication Without Loss -- Merged signals retain references to all original sources
5. Actionable Output -- Ranked list with clear tickers, direction, and time horizon for each idea
6. Idempotent Aggregation -- Re-ingesting the same evidence never inflates scores; agreement credit requires distinct skills AND distinct events
7. Strict Score Validation -- NaN/infinity, negative, and out-of-range scores are rejected and reported, never converted into maximum confidence