Structured web research framework for AI agents. Teaches your agent to conduct multi-source research, synthesize findings into actionable briefs, maintain a research library, and track evolving topics over time. Use when you need market research, competitor analysis, topic deep-dives, or ongoing monitoring of trends and news. Works with any agent that has web search capabilities.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
A direct command skips the review prompt. Inspect the source before running it.
The command stays on one line. Scroll horizontally to inspect it before copying.
Prefer a local copy? Download the files currently available to SkillsMP.
Showing SKILL.md
SKILL.md
Source instructions · Read-only preview
name
research-assistant
version
1.0.0
description
Structured web research framework for AI agents. Teaches your agent to conduct multi-source research, synthesize findings into actionable briefs, maintain a research library, and track evolving topics over time. Use when you need market research, competitor analysis, topic deep-dives, or ongoing monitoring of trends and news. Works with any agent that has web search capabilities.
By The Agent Ledger — Subscribe for more agent skills and premium blueprints.
Turn your AI agent into a structured research analyst. No more copy-pasting search results — your agent learns to plan research, cross-reference sources, synthesize findings, and maintain an evolving knowledge base.
What This Skill Does
Structured Research Protocol — Plan → Search → Evaluate → Synthesize → Store
Research Brief Format — Consistent, scannable output for every research task
Source Quality Scoring — Teaches your agent to evaluate and rank sources
Research Library — Organized storage for findings that persists across sessions
Topic Monitoring — Track evolving topics with periodic re-research
research/
├── README.md ← Library index (auto-maintained)
├── briefs/ ← Completed research briefs
├── monitoring/ ← Topics being tracked over time
└── templates/ ← Custom research templates (optional)
Step 2: Create research/README.md
# Research Library## Recent Briefs
<!-- Agent maintains this list automatically -->
## Monitored Topics
<!-- Topics being tracked with periodic updates -->
## Quick Stats- Total briefs: 0
- Active monitors: 0
- Last research: never
Step 3: Add to Your Agent Instructions
Add to your AGENTS.md, SOUL.md, or system prompt:
## Research Protocol
When asked to research something:
1. Read `research/README.md` for existing work on this topic
2. Follow the Research Protocol below
3. Save the brief to `research/briefs/YYYY-MM-DD-<slug>.md`
4. Update `research/README.md` index
Research Protocol
Phase 1: Plan
Before searching, define:
Research Question — What specifically are we trying to answer?
Scope — Broad survey or deep dive? Time range? Geography?
Sources Needed — How many independent sources for confidence?
Output Format — Brief, comparison table, recommendation, or raw dump?
Phase 2: Search & Collect
Execute searches with these principles:
Multiple queries — Rephrase the question 2-3 ways for coverage
Source diversity — Mix official sources, industry analysis, community discussion
Recency check — Note publication dates; flag anything older than 6 months
Contradiction hunting — Actively look for sources that disagree
Phase 3: Evaluate Sources
Score each source on:
Factor
Weight
Criteria
Authority
High
Known publication? Expert author? Official source?
Recency
High
Published within relevant timeframe?
Specificity
Medium
Concrete data/examples vs. vague claims?
Corroboration
Medium
Do other sources confirm this?
Bias Check
Low
Obvious commercial or political motivation?
Drop sources scoring poorly on Authority + Recency. Flag but keep sources with clear bias if they contain unique data.
Phase 4: Synthesize
Structure findings into a Research Brief (see format below). Key rules:
Lead with the answer — Don't bury the conclusion
Quantify when possible — Numbers > adjectives
Flag uncertainty — Clearly mark low-confidence claims
Note gaps — What couldn't you find? What needs deeper research?
Phase 5: Store & Index
Save to research/briefs/YYYY-MM-DD-<slug>.md and update the library index.
Research Brief Format
# [Research Question]**Date:** YYYY-MM-DD
**Requested by:** [context]
**Confidence:** High / Medium / Low
**Staleness risk:** [when this research might become outdated]
## TL;DR
[2-3 sentence executive summary with the key finding]
## Key Findings### Finding 1: [Title]
[Detail with supporting evidence]
- Source: [name] ([link]) — [date]
- Confidence: High/Medium/Low
### Finding 2: [Title]
[Detail with supporting evidence]
- Source: [name] ([link]) — [date]
- Confidence: High/Medium/Low
[Continue as needed]
## Data Points
| Metric | Value | Source | Date |
|--------|-------|--------|------|
| [key stat] | [value] | [source] | [date] |
## Contradictions & Caveats- [Any conflicting information found]
- [Limitations of this research]
## Knowledge Gaps- [What we couldn't find]
- [What needs deeper investigation]
## Sources1. [Full citation with URL and access date]
2. [...]
## Recommendations- [Actionable next steps based on findings]
---
*Research conducted by AI agent using the Research Assistant skill by [The Agent Ledger](https://theagentledger.com)*
Topic Monitoring
For topics that evolve over time (market trends, competitor moves, regulatory changes):
Setup a Monitor
Create research/monitoring/<topic-slug>.md:
# Monitoring: [Topic Name]**Started:** YYYY-MM-DD
**Frequency:** [daily / weekly / bi-weekly]
**Search queries:**- "[query 1]"
- "[query 2]"
**Alert triggers:** [what constitutes a notable change]
## Updates### YYYY-MM-DD- [What changed since last check]
- [New data points]
- [Assessment: significant / minor / no change]
Periodic Check Protocol
When running a monitoring check:
Read the monitor file for context and previous findings
Run the saved search queries
Compare new results against last update
Only log if something changed or enough time has passed
Flag significant changes prominently
Integration with Heartbeats/Cron
Add to your HEARTBEAT.md or set up a cron:
## Research Monitors
Check research/monitoring/ for topics due for refresh.
Only check topics whose frequency interval has elapsed.
Increase depth level; add "find primary sources" to prompt
Too many sources, no synthesis
Emphasize Phase 4; ask for "TL;DR first, then details"
Findings are outdated
Add recency filter: "only sources from last 6 months"
Agent dumps raw search results
Reinforce the brief format; remind to synthesize, not copy
Monitor checks too frequent
Adjust frequency in monitor file; use staleness risk to guide
Research library getting messy
Run periodic cleanup: archive briefs older than 6 months
Integration with Other Agent Ledger Skills
Solopreneur Assistant — Feed research into business decisions and opportunity scoring
Daily Briefing — Include monitor alerts in daily briefings
Project Tracker — Link research briefs to project milestones
Memory OS — Store key research insights in long-term memory
Part of the free skill library by The Agent Ledger. Subscribe for premium blueprints, playbooks, and the complete agent configuration guide.
License: CC-BY-NC-4.0
DISCLAIMER: This skill was created by an AI agent. It is provided "as is" for informational and educational purposes only. It does not constitute professional, financial, legal, or technical advice. Review all generated files before use. The Agent Ledger assumes no liability for outcomes resulting from use. Use at your own risk.
