skills.rest
MarcoBolsa
MarcoBolsaCOMMUNITY
Explore all skills by this author →
GitHub

tech-search

Executes multi-phase deep tech research using parallel web search workers and saves structured reports.

Updated May 13, 2026
Search Tags:#documentation#research#web search#synthesis#tech research#query decomposition#parallel workers
One-click install
npx skills add https://github.com/MarcoBolsa/motor-cotacoes --skill tech-search-marcobolsa
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: tech-search
Source: https://github.com/MarcoBolsa/motor-cotacoes/tree/main/.claude/skills/tech-search
Command: npx skills add https://github.com/MarcoBolsa/motor-cotacoes --skill tech-search-marcobolsa

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Conducting thorough technical research requires running many searches, reading sources, and synthesizing findings manually. This Skill automates the entire pipeline from query decomposition to a documented research report. ## Core Features & Use Cases - Automated Research Pipeline: Decomposes a query into 5-7 sub-queries, dispatches parallel Haiku workers for web search and deep page reads, then evaluates coverage across up to 2 search waves. - Structured Documentation: Saves results to docs/research/{date}-{slug}/ with an index, original query, research prompt, full report, and recommendations. - Use Case: Ask "React Server Components vs Client Components" and receive a synthesized comparison report with key findings, code examples, expert quotes, and a decision matrix, without writing any code. ## Quick Start Run /tech-search followed by your research question, for example asking it to research the tradeoffs between PostgreSQL and MongoDB for a new project.

Frequently Asked Questions about tech-search

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a deep tech research query with this skill?▼

Invoke /tech-search followed by your question in quotes, such as "/tech-search 'React Server Components vs Client Components'". The skill auto-clarifies the query, decomposes it into sub-queries, searches in parallel, and saves a full report to docs/research/.

What tools does tech-search use for web research?▼

It prefers the Exa MCP for search and Context7 MCP for library documentation when available, falling back to WebSearch otherwise. Deep page reads always use WebFetch with a structured extraction prompt, with no external scripts or dependencies.

Can tech-search implement code or create agents from research?▼

No. The skill explicitly vetoes implementation requests and redirects them to @pm or @dev agents. It produces documentation only, and code examples in reports are for reference, never production artifacts.

Where does tech-search save research output files?▼

All output is written to docs/research/{YYYY-MM-DD}-{slug}/ containing a README index, the original query, the generated research prompt, the full report, and recommendations. Writing outside docs/research/ is blocked.

What happens when web searches return no results?▼

If all search waves return zero results, the skill stops and reports that no results were found, suggesting you reformulate the query or check connectivity. It requires at least one successful worker result to proceed.

Related Skills

Explore More Education & Research Skills→
SynkraAISynkraAI
3.1k

tech-search

Researches technical questions via parallel web search workers and saves structured reports to docs/research.

Official
Advanced
psd401psd401
2

deepen-plan

Validate plan sections against best practices and annotate with research notes.

Official
Advanced
Jimmy-JungJimmy-Jung

research

Coordinate parallel research across official docs, web searches, and codebases.

Community
Advanced
awood45awood45
11

quick-research

Finds sourced, concise answers to focused factual questions using web search and fetch.

Community
Basic