Collector Hand

What It Does

Collector Hand is an OSINT-grade intelligence collection agent that monitors any target (company, person, technology, market, topic) 24/7, building a living knowledge graph and detecting changes over time. You give it a target. It watches. When something important changes—leadership shift, funding round, product launch, sentiment swing—it alerts you.

Key Features

Continuous monitoring: Hourly, daily, or weekly collection sweeps
Change detection: Compares each run against previous state, flags new/changed/disappeared entities
Knowledge graph construction: Entities (people, companies, products, events) + relationships
Sentiment tracking: Monitors positive/negative/neutral trends over time
Multi-focus analysis: Market intelligence, business intelligence, competitor analysis, person tracking, technology monitoring
OSINT methodology: 5-phase collection cycle (planning, collection, processing, analysis, dissemination)

Activation

# Activate Collector Hand with a target
openfang hand activate collector --settings "target_subject=OpenAI,focus_area=business"

# Monitor a competitor
openfang hand activate collector --settings "target_subject=CompetitorCo,focus_area=competitor,update_frequency=daily"

# Track a person
openfang hand activate collector --settings "target_subject=John Doe,focus_area=person"

Configuration Settings

target_subject

text

What to monitor: company name, person, technology, market, or topic

collection_depth

select

default:"deep"

How deep to dig on each cycle:

surface: Headlines only
deep: Full articles + sources (default)
exhaustive: Multi-hop research

update_frequency

select

default:"daily"

How often to run collection sweeps:

hourly: Every hour
every_6h: Every 6 hours
daily: Daily (default)
weekly: Weekly

focus_area

select

default:"general"

Lens through which to analyze collected intelligence:

market: Market intelligence (trends, players, reports)
business: Business intelligence (revenue, partnerships, strategy)
competitor: Competitor analysis (product launches, pricing, reviews)
person: Person tracking (publications, job changes, social activity)
technology: Technology monitoring (releases, adoption, benchmarks)
general: General intelligence (default)

alert_on_changes

toggle

default:"true"

Publish an event when significant changes are detected (critical or important level).

report_format

select

default:"markdown"

Output format for intelligence reports:

markdown: Markdown report (default)
json: JSON data
html: HTML report

max_sources_per_cycle

select

default:"30"

Maximum number of sources to process per collection sweep:

10 sources
30 sources (default)
50 sources
100 sources

track_sentiment

toggle

default:"false"

Analyze and track sentiment trends over time (positive/negative/neutral classification).

Required Tools

Collector Hand requires access to these tools (all built-in):

shell_exec — Platform detection
file_read, file_write, file_list — Knowledge base persistence
web_fetch, web_search — Source collection
memory_store, memory_recall — State persistence
schedule_create, schedule_list, schedule_delete — Continuous monitoring
knowledge_add_entity, knowledge_add_relation, knowledge_query — Knowledge graph
event_publish — Critical change alerts

System Prompt Overview

Collector Hand operates in 7 phases:

Platform Detection & State Recovery

Detects OS, loads previous collection state, reads configuration, queries knowledge graph for existing entities.

Schedule & Target Initialization

Creates collection schedule based on update frequency. Identifies target type (company/person/technology/market). Builds initial query set tailored to target and focus area.

Source Discovery & Query Construction

Builds targeted search queries based on focus area (market intelligence, business intelligence, competitor analysis, etc.). Adds temporal queries for recency.

Collection Sweep

Executes queries, fetches full content, extracts entities (people, companies, products, dates, numbers, events). Tags each data point with source, timestamp, confidence level, and relevance score.

Knowledge Graph Construction

Adds entities and relationships to the knowledge graph. Entity types: Person, Company, Product, Event, Number. Relation types: works_at, founded, invested_in, partnered_with, competes_with, launched, acquired.

Change Detection & Delta Analysis

Compares current collection against previous snapshot. Identifies new entities, changed attributes, new relationships, disappeared entities. Scores each change by significance (critical/important/minor). Publishes event if critical changes found.

Report Generation

Generates intelligence report in configured format. Saves updated knowledge base to collector_knowledge_base.json. Updates dashboard statistics.

Usage Examples

Competitor Monitoring

openfang hand configure collector \
  --set target_subject="CompetitorCo" \
  --set focus_area="competitor" \
  --set update_frequency="daily" \
  --set alert_on_changes="true"

openfang hand activate collector

Collector Hand will run daily, monitoring:

Product launches
Pricing changes
Customer reviews
Hiring patterns
Funding announcements
Partnership news

When something critical happens (e.g., major product launch), it publishes an alert event.

Market Intelligence

openfang hand configure collector \
  --set target_subject="AI agent frameworks" \
  --set focus_area="market" \
  --set collection_depth="exhaustive" \
  --set update_frequency="weekly"

openfang hand activate collector

Collector Hand will run weekly, gathering:

Market size and growth trends
Key players and their market share
Industry reports
Regulatory changes
Technology adoption metrics

Person Tracking

openfang hand configure collector \
  --set target_subject="Satya Nadella" \
  --set focus_area="person" \
  --set update_frequency="daily"

openfang hand activate collector

Collector Hand will monitor:

Public talks and interviews
Publications and papers
Job changes
Social media activity
News mentions

Change Detection

Collector Hand classifies each change by significance:

Critical (Immediate Alert)

Leadership change (CEO, CTO, board)
Acquisition or merger
Major funding round (>$10M)
Product discontinuation
Legal action or regulatory issue

Important (Include in Next Report)

New product launch
New partnership or integration
Hiring surge (>5 roles)
Pricing change
Competitor move
Major customer win/loss

Minor (Note in Report)

Blog post or press mention
Minor update or patch
Social media activity spike
Conference appearance
Individual job posting

Dashboard Metrics

Collector Hand tracks four key metrics:

Data Points

Total data points collected across all cycles.

Entities Tracked

Unique entities in the knowledge graph.

Reports Generated

Total intelligence reports produced.

Last Update

Timestamp of the most recent collection.

View in the dashboard at http://localhost:4200/hands/collector.

Sentiment Tracking

When track_sentiment is enabled, Collector Hand classifies each source:

Strong positive (+2): “Company wins major award”
Mild positive (+1): “Steady growth continues”
Neutral (0): “Company releases Q3 report”
Mild negative (-1): “Faces increased competition”
Strong negative (-2): “Major data breach disclosed”

Tracks rolling average over last 5 collection cycles to detect trends.

Report Format

# Intelligence Report: [target_subject]
**Date**: YYYY-MM-DD | **Cycle**: N | **Sources Processed**: X

## Key Changes Since Last Report
- [CRITICAL] OpenAI appoints new Chief Scientist (source)
- [IMPORTANT] Launched GPT-5 with 10T parameters (source)

## Intelligence Summary
[2-3 paragraph synthesis of collected intelligence]

## Entity Map
| Entity | Type | Status | Confidence |
|--------|------|--------|------------|
| Greg Brockman | Person | CTO | high |
| GPT-5 | Product | Launched | high |

## Sources
1. [TechCrunch](url) — confidence: high — extracted: leadership change
2. [OpenAI Blog](url) — confidence: high — extracted: product launch

## Sentiment Trend (if enabled)
Positive: 65% | Neutral: 25% | Negative: 10% | Trend: up (+15% vs last cycle)

Knowledge Graph Structure

Collector Hand builds a living knowledge graph:

{
  "entities": [
    {
      "entity_id": "openai_001",
      "name": "OpenAI",
      "type": "company",
      "attributes": {
        "industry": "AI",
        "founded": "2015",
        "ceo": "Sam Altman"
      },
      "sources": ["https://openai.com", "https://techcrunch.com/..."],
      "first_seen": "2025-01-01T00:00:00Z",
      "last_seen": "2025-03-07T00:00:00Z",
      "confidence": "high"
    }
  ],
  "relations": [
    {
      "source_entity": "sam_altman_001",
      "relation": "works_at",
      "target_entity": "openai_001",
      "attributes": {"role": "CEO", "since": "2019"},
      "source": "https://openai.com/about",
      "confidence": "high"
    }
  ]
}

Best Practices

Collector Hand will never fabricate intelligence. Every claim is sourced. It distinguishes facts from analysis/speculation.

For competitor monitoring, use daily frequency. For market intelligence, weekly is often sufficient.

Enable alert_on_changes for time-sensitive targets (competitors, key accounts). Disable it for historical trend analysis.

Combine Collector Hand with Researcher Hand: Collector finds the signals, Researcher deep-dives on the important ones.

Advanced Configuration

Custom Query Sets

Manually trigger Collector Hand with custom collection guidance:

openfang chat collector
> "Focus on OpenAI's partnerships and integrations announced in the last 30 days."

Multi-Target Monitoring

Run multiple Collector Hands in parallel:

openfang hand activate collector --id competitor1 --settings "target_subject=CompetitorA"
openfang hand activate collector --id competitor2 --settings "target_subject=CompetitorB"
openfang hand activate collector --id competitor3 --settings "target_subject=CompetitorC"

Get Started

Core Concepts

Autonomous Hands

Configuration

Guides

Security

Deployment

What It Does

Key Features

Activation

Configuration Settings

Required Tools

System Prompt Overview

Usage Examples

Competitor Monitoring

Market Intelligence

Person Tracking

Change Detection

Dashboard Metrics

Data Points

Entities Tracked

Reports Generated

Last Update

Sentiment Tracking

Report Format

Knowledge Graph Structure

Best Practices

Advanced Configuration

Custom Query Sets

Multi-Target Monitoring

Next Steps

Lead Hand

Predictor Hand

Build docs developers (and LLMs) love

Get Started

Core Concepts

Autonomous Hands

Configuration

Guides

Security

Deployment

​What It Does

​Key Features

​Activation

​Configuration Settings

​Required Tools

​System Prompt Overview

​Usage Examples

​Competitor Monitoring

​Market Intelligence

​Person Tracking

​Change Detection

​Dashboard Metrics

Data Points

Entities Tracked

Reports Generated

Last Update

​Sentiment Tracking

​Report Format

​Knowledge Graph Structure

​Best Practices

​Advanced Configuration

​Custom Query Sets

​Multi-Target Monitoring

​Next Steps

Lead Hand

Predictor Hand

Build docs developers (and LLMs) love

What It Does

Key Features

Activation

Configuration Settings

Required Tools

System Prompt Overview

Usage Examples

Competitor Monitoring

Market Intelligence

Person Tracking

Change Detection

Dashboard Metrics

Sentiment Tracking

Report Format

Knowledge Graph Structure

Best Practices

Advanced Configuration

Custom Query Sets

Multi-Target Monitoring

Next Steps