spotlight-handoff

name	spotlight-handoff
description	The handoff phase of a data-detective investigation — translate finalized data-side findings into Spotlight-ready briefs and pass leads to the spotlight orchestrator for external OSINT amplification. Each candidate finding becomes a spotlight-brief.md with named entities, URLs already gathered, what data-detective established (with evidence card refs), and suggested external angles. The skill also generates a spotlight-handoff.md ranking sheet so the user can pick which leads warrant Spotlight's web-based investigation. Use in Phase 6 of a data-detective investigation, after Gate 1 approval. Triggers on prompts mentioning passover, handoff to spotlight, generate OSINT briefs, what should we research externally, or "take these findings and dig deeper online."
license	MIT
metadata	{"parent":"data-detective","step":"P6_passover","pattern":"data → web OSINT handoff","inputs":"finalized findings.json + evidence-map.json","outputs":"spotlight-handoff.md (ranking) + spotlight-briefs/*.md (one per chosen lead)"}

data-detective-passover

The handoff cycle: data-side findings → Spotlight briefs → external OSINT.

When to use

Called from the data-detective orchestrator at Phase 6, after the user approves Gate 1 (case/summary.md).

This is the passover cycle: data-detective has found anomalies in the corpus and verified them via the adversarial fact-checker. Now we hand off to Spotlight for the external-world OSINT — court records (PACER), corporate filings (Delaware DOC, OpenCorporates), social media, news archives, FOIA traces, business networks.

The combination — data-side audit trail + Spotlight-side web OSINT — is richer than either alone.

Two-step protocol

Step 1 — Generate the candidates ranking sheet

For each verified finding plus each unchased thread from the detector CSVs, emit a row in case/spotlight-handoff.md:

# Candidates for Spotlight OSINT amplification
Generated: <ISO timestamp>

| Rank | ID | One-line claim | OSINT yield | Effort | Suggested angle |
|---|---|---|---|---|---|
| 1 | OS-001 | <claim> | high | medium | <angle> |
| 2 | OS-002 | ... | ... | ... | ... |

Rank by:

• OSINT yield: how much external context would amplify this? (named individuals + corporate connections + political stakes = high)
• Effort: how much OSINT work is realistic? (one court case to pull = low; full network mapping = high)
• Newsworthiness: how strong is the story if external evidence lands?

Present the table to the user. User picks which to pursue.

Step 2 — Generate Spotlight briefs for picked candidates

For each candidate the user picks, emit case/spotlight-briefs/<id>.md:

# Spotlight investigation brief — OS-<NNN>

**Lead.** <one-sentence claim from data-detective>

**Source.** Hand-off from data-detective finding C-<NNN> (see ../findings-report.md for full evidence chain).

**Confidence level (data-side).** verified | partially_verified | medium | high

## Named entities for OSINT

### Priority targets
1. **<Person/Org name>** — <one-line role>. Known so far: <bullets>. Open OSINT questions: <bullets>.
2. ...

### Secondary targets
- <name> — <role>

## URLs already in the case-trace (start here)

- <lda.senate.gov public filing URL>
- <courtlistener URL if relevant>
- <entity-own website if relevant>
- <archived snapshot URLs from case/external/factcheck/>

## What data-detective established

<3-5 bullet summary with evidence-card refs (filing UUIDs / press_ids / FARA registration numbers)>

## What Spotlight should pursue (suggested angles)

- Corporate registrations: <state, search terms>
- Court records: <PACER docket numbers if any> + appeals
- Social media / web: <handles, profile URLs>
- Network mapping: <funded by, board memberships, prior employers>
- FOIA: <agencies that should have records>
- Adjacent press coverage: <topical search terms>

## Out of scope

<things data-detective intentionally did not pursue and Spotlight should not feel obligated to>

Step 3 — Spawn Spotlight on the chosen brief

Skill(spotlight) with brief=case/spotlight-briefs/<id>.md

Spotlight runs its own preflight (config, search library check), accepts the brief at Phase 1 (skipping the conversational brief stage since the data-detective brief is detailed enough), and proceeds through its own cycles. Its output (web-archived sources, social-media-intelligence outputs, etc.) joins case/external/spotlight-<id>/ to merge with the data-detective trace.

Why this works

• Data-detective is great at structured-records anomaly detection. It can find that "$180M of LDA filings come from one Delaware shell named after a 'Loc Nation'" in seconds against a 8.6 GB corpus.
• Spotlight is great at structured external investigation. It can pull PACER, archive Reddit / Facebook profile pages, map shell-LLC networks via Delaware DOC, and verify against open sanctions lists.
• Neither alone tells the whole story. Sullivan lobbying Apollo while Warren attacks Apollo is a data-detective insight; what Apollo and the people moving between them are doing in the world is a Spotlight insight.

The passover protocol formalizes the handoff so the work composes cleanly: data-detective's findings become Spotlight's brief at Phase 1.

Reverse handoff (optional, Phase 7+)

Once Spotlight's external investigation completes, its findings can be brought BACK into data-detective as a new detector seed. Example: Spotlight identifies a new actor name → data-detective greps the corpus for that name → finds N additional filings → loops back through P3.

Anti-patterns

• No automatic spawn without user choice. Always present spotlight-handoff.md and let the user pick. Spotlight cycles cost real time/tokens; the user picks which leads to spend on.
• No briefs without an evidence card link. Every Spotlight brief must reference at least one data-detective evidence card by ID so the data-side audit trail is preserved.
• No briefs that re-investigate what data-detective already established. The brief asks Spotlight to amplify, not to re-verify.

Composes with

• Input ← data-detective orchestrator (Phase 5 synthesis output: findings-report.md, evidence-map.json)
• Output → Skill(spotlight) with brief
• Reverse output → seeds for next data-detective cycle