parallel-research
Skill0 starsChanged 6 months agoMITView on GitHub
Multi-agent parallel research methodology. 3-5 agents investigate different perspectives of a topic simultaneously for comprehensive coverage with minimal overlap. 4x speedup vs sequential research.
6 other projects use a version of this.
Adopted by
Most recently taken up by
- tymoshenko-dmytro/marketing-agent-kittheir own version, first seen Oct 21, 2026
- simonstrumse/vibelabs-skillstheir own version, first seen Oct 15, 2026
- tymoshenko-dmytro/claude-marketing-skillstheir own version, first seen Oct 13, 2026
- jdurey/skillstheir own version, first seen Jul 19, 2026
- Astro-Han/skillstheir own version, first seen Jul 12, 2026
- Jay-523/agent-skillstheir own version, first seen Mar 28, 2026
Other versions6 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 6
- Jay-523/agent-skills★ 17same name
- Astro-Han/skills★ 12same name
- tymoshenko-dmytro/claude-marketing-skills★ 2same name
- tymoshenko-dmytro/marketing-agent-kit★ 2same name
- jdurey/skills★ 0same name
- simonstrumse/vibelabs-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository17 other files it gives its agents.
- blog-banner-creationSkill
- daily-blog-draftSkill
- image-generationSkill
- script-to-speechSkill
- company-deep-diveSkill
- email-verificationSkill
- funnel-playbookSkill
- google-ads-auditSkill
- landing-page-funnelSkill
- linkedin-content-image-typesSkill
- viral-comment-strategySkill
- autonomous-task-reliabilitySkill
- build-your-brainSkill
- diagram-generatorSkill
- expert-councilSkill
- liaclSkill
- rate-limit-safeguardsSkill
The file
We split this skill into the rules it gives an agent
6 rules. 0 are common, 6 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 6 edited versions we read keep these.
6
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 50
Ensure perspectives are complementary, not redundant
- Distinctive: line 66
All researchers work in parallel:
- Distinctive: line 82
Read all angle-specific reports
- Distinctive: line 85
Create unified answer that honors all angles
- Distinctive: line 94
Verify coverage and quality:
- Distinctive: line 160
Error handling: If one angle fails, others still provide value
“Common” is measured: kept by at least 6 in 10 of the 6 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Parallel Research Skill
A multi-perspective research methodology where 3-5 agents (or research threads) simultaneously investigate different angles of a topic. By dividing perspectives (not just workload), you achieve comprehensive coverage with minimal overlap and diverse insights.
When to Use
Invoke when:
- Complex, multi-perspective questions need answering
- Comprehensive topic coverage required quickly
- Limited time for deep research (parallel > sequential)
- Need diverse viewpoints on same topic
- Research question has multiple valid angles
Do not use when:
- Single clear domain (a single specialist is more efficient)
- Need sequential depth on one angle
- Research requires dependent discoveries (each step needs prior)
- Simple factual lookup
Prerequisites
Roles Needed:
- External researcher -- Industry context, market data, published sources
- Pattern analyst -- Structural/architectural patterns
- Historical researcher -- Prior art, evolution, context
- Risk analyst -- Security, risk, vulnerability angle
- Synthesizer -- Consolidates all perspectives
Context Needed:
- Clear research question
- Angle assignments (who researches what aspect)
- Synthesis criteria (what makes a good combined answer)
Procedure
Step 1: Angle Assignment (~5 minutes)
Divide the research into non-overlapping angles:
- Decompose question into 3-5 distinct perspectives
- Assign each perspective to an appropriate researcher
- Ensure perspectives are complementary, not redundant
- Brief each researcher on their specific angle
Example Angles:
- Technical implementation
- Industry precedent
- Security implications
- User experience impact
- Historical context
Deliverable: Angle assignment matrix
Step 2: Parallel Investigation (~30-60 minutes, all run simultaneously)
All researchers work in parallel:
- Each focuses ONLY on their assigned angle
- Gather evidence, examples, insights
- Note confidence levels
- Flag cross-cutting discoveries for synthesis
- Produce angle-specific report
Deliverable: 3-5 specialist research reports
Step 3: Synthesis (~15-20 minutes)
Consolidate perspectives:
- Read all angle-specific reports
- Identify cross-perspective themes
- Note contradictions or tensions
- Create unified answer that honors all angles
- Produce confidence-weighted synthesis
Deliverable: Synthesized research findings
Step 4: Quality Check (~5 minutes)
Verify coverage and quality:
- Calculate overlap percentage (target: <15%)
- Assess coverage completeness
- Rate overall quality
- Document any gaps
Deliverable: Quality assessment
Parallelization
Can run in parallel:
- Step 2 (the heart of this method) -- All research happens simultaneously
- This is where the 4x speedup comes from
Must be sequential:
- Step 1 before 2 (assignments needed)
- Step 2 before 3 (reports needed for synthesis)
- Step 3 before 4 (synthesis needed for quality check)
Success Indicators
- 3-5 specialist research reports completed
- <15% overlap between reports (truly different perspectives)
- Synthesis addresses all angles coherently
- Quality rating 8.5+ across all reports
- Comprehensive coverage achieved (no major gaps)
- 3-4x faster than sequential research would have been
Example
Scenario: Research "AI agent coordination patterns" across perspectives
Step 1 (Assign):
- External researcher: Industry examples and academic papers
- Pattern analyst: Structural patterns and architectures
- Historical researcher: Historical approaches and evolution
- Risk analyst: Security implications of multi-agent systems
Step 2 (Research) - all parallel:
External: Found 7 industry implementations, 3 papers
Pattern: Identified 4 core patterns (hub, mesh, hierarchy, swarm)
Historical: Traced 15-year evolution from simple scripts
Risk: Flagged 3 attack vectors, 2 mitigations
Step 3 (Synthesize):
Unified finding: "Effective coordination requires..."
Cross-theme: All reports emphasized trust/verification
Tension: Hierarchy (efficient) vs. Mesh (resilient)
Step 4 (Quality):
Overlap: 8% (excellent - truly different perspectives)
Coverage: Comprehensive
Quality: 9.2/10 average
Speed: 4x faster than sequential
Result: Comprehensive answer from 4 perspectivesKey Insights
- Divide PERSPECTIVES, not just workload -- this is what makes parallel research better than simply splitting work
- The "<15% overlap" metric validates truly different angles
- Error handling: If one angle fails, others still provide value
- Scale: Can extend to 6-8 researchers for very complex topics
- When NOT to use: Single-domain questions where a single specialist is more efficient
Built by PureBrain -- AI-powered marketing operations.
Want the full suite of 183+ production-tested skills? Start your trial ->
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

