emaduneme Information Gathering & Research Skill
What this skill does
Equips journalists with structured workflows and tools to:
- Monitor beats via RSS feeds and keyword alerts
- Aggregate and summarize recent coverage on a topic
- Build source lists (experts, officials, affected communities)
- Find primary sources: public records, datasets, government documents
- Generate a research brief before field reporting
Workflow
Step 1 — Define the research scope
Clarify:
- Topic: What are you researching?
- Purpose: Background context? Source-finding? Breaking news monitoring?
- Time horizon: Current (past 7 days) / Recent (past 3 months) / Historical?
- Geography: Local, national, international?
Step 2 — Aggregate current coverage
Run web_search using multiple query types:
"[topic]" news [current month year]
"[topic]" site:gov OR site:edu OR site:org
"[topic]" data OR statistics OR report OR study
"[topic]" "according to" OR "spokesperson" OR "official said"
For ongoing beat monitoring, use the RSS monitor script:
python scripts/rss_monitor.py --feeds feeds.json --keywords "keyword1,keyword2" --days 7
Step 3 — Find primary sources
Public records by category:
| Source | Type | URL |
|---|---|---|
| data.gov | Federal datasets | data.gov |
| PACER | Federal court records | pacer.gov |
| SEC EDGAR | Corporate filings | sec.gov/edgar |
| ProPublica Nonprofit Explorer | 990 tax forms | projects.propublica.org/nonprofits |
| Google Scholar | Academic research | scholar.google.com |
| Bellingcat Toolkit | OSINT/investigation | bellingcat.com |
Use web_search: "[organization]" site:sec.gov or "[topic]" filetype:pdf site:gov
Run the source finder for nonprofit/corporate lookups:
python scripts/source_finder.py --type nonprofit --name "Organization Name"
python scripts/source_finder.py --type corporate --name "Company Name"
Step 4 — Build a source list
Every story needs sources across these categories:
- Official: Spokespeople, government agencies, institutions
- Expert: Academics, researchers, think tanks (Google Scholar, university directories)
- Affected community: People directly experiencing the issue
- Opposing voices: Critics, advocates with a different view
- Documentary: Reports, datasets, public records
web_search query for experts: "[topic]" professor OR researcher "[relevant field]" contact
Step 5 — Produce a research brief
RESEARCH BRIEF: [Topic]
Date: [Today]
Prepared for: [Story/Reporter]
BACKGROUND (3-5 sentences):
[Context, history, current status]
KEY FACTS:
- [Fact | Source URL]
- [Statistic | Source URL]
PRIMARY SOURCES FOUND:
- [Document name | URL | Why relevant]
EXPERT SOURCES:
- [Name | Affiliation | Why relevant | Contact if found]
COVERAGE GAPS:
- [What has NOT been reported]
SUGGESTED NEXT STEPS:
- [FOIA request for X]
- [Interview Y]
- [Verify Z with dataset]
Python Scripts
scripts/rss_monitor.py
Monitors a configurable list of RSS feeds for keyword matches. Outputs a digest of matched stories with headlines, sources, and links.
# First, populate feeds.json with your RSS URLs (see feeds.json.example)
python scripts/rss_monitor.py --feeds feeds.json --keywords "eviction,housing" --days 3
python scripts/rss_monitor.py --feeds feeds.json --keywords "budget cuts" --output digest.txt
scripts/source_finder.py
Searches public databases (ProPublica Nonprofit API, OpenCorporates) for organizational records.
python scripts/source_finder.py --type nonprofit --name "Red Cross"
python scripts/source_finder.py --type news --query "water contamination" --days 30
Notes for the AI
- Always label facts as "needs verification" unless they come from a primary source you have directly examined
- Government (.gov), academic (.edu), and established nonprofit sites have higher baseline credibility than news aggregators
- Present the source URL alongside every fact pulled from web search
- For ongoing beats, recommend setting up the RSS monitor rather than repeating manual searches
- When generating a source list, do not invent contact information — find it or flag it as missing