Przejdź do treści
Biblioteka skilli
Programowanie · Research

deep-research

/deep-research

Bada nieznane technologie i zgodność frameworków oraz bibliotek, weryfikuje sygnatury API i zestawia dowody techniczne.

Licencja
MIT
Ostatnia zmiana repozytorium
13.11.2026

Ocena i bezpieczeństwo

90/ 100Klasa: dobry
Format25 / 25
Frontmatter z nazwą i opisem, poprawna nazwa, opis do 1024 znaków, nazwa zgodna z folderem.
Odkrywalność25 / 25
Opis mówi, kiedy użyć skilla, ma sensowną długość i jest w 3. osobie.
Treść25 / 35
Co najmniej 150 słów i do 500 linii, nagłówki, listy kroków, przykłady albo kod.
Higiena15 / 15
Bez zaślepek typu TODO, bez ścieżek Windows, z licencją.
Ryzykobrak sygnałów w skanie
Pliki w paczce
2
Skrypty
0
Komentarze HTML
0
Adresy URL w SKILL.md
0
Przestarzałość

Brak sygnałów. Repozytorium zmienione 13.11.2026, bez odwołań do wycofanych modeli i API.

Licencja

MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.

Co robi i kiedy

Bada nieznane technologie i zgodność frameworków oraz bibliotek, weryfikuje sygnatury API i zestawia dowody techniczne.

Oryginalny opis (EN)

Use when investigating unfamiliar technologies, researching framework/library compatibility, verifying API signatures, evaluating technical alternatives, cross-referencing security advisories, or synthesizing technical evidence.

Jak działa

  1. Purpose
  2. Activation criteria
  3. Non-activation guidelines
  4. Required context and inputs
  5. Operating procedure
  6. Technical and engineering standards
  7. Verification and evidence requirements
  8. Expected outputs and deliverables
  9. Risks, boundaries and failure modes
  10. Progressive disclosure references
Sekcje z nagłówków SKILL.md, w oryginale · 84 linii, 713 słów
Podgląd SKILL.md
---
name: "deep-research"
description: "Use when investigating unfamiliar technologies, researching framework/library compatibility, verifying API signatures, evaluating technical alternatives, cross-referencing security advisories, or synthesizing technical evidence."
---

# Deep Research — Technical Investigation and Authoritative Evidence Synthesis

> Generated from `specs/skills/deep-research.md` by `scripts/generate-antigravity-skills.ts`. Edit the canonical specification, not this file.

## Purpose

The Deep Research skill provides a disciplined methodology for investigating unfamiliar technologies, complex APIs, library compatibility, and technical trade-offs. It ensures that engineering decisions are grounded in verified, authoritative primary sources, current version specifications, and explicit trade-off analyses, preventing reliance on stale forum advice or hallucinated capabilities.

## Activation criteria

Activate this skill when:

- Investigating an unfamiliar framework, library, API, or database technology.
- Resolving dependency compatibility conflicts, peer dependency warnings, or breaking changes.
- Evaluating multiple competing technology choices for a new project or subsystem.
- Assessing security advisories, CVE reports, or regulatory requirements.
- Determining whether an official platform API supports a requested feature.

## Non-activation guidelines

Do not activate this skill when:

- Implementing already-decided features using existing codebase conventions.
- Debugging local syntax errors or simple runtime exceptions (use Advanced Debugging).
- Authoring component unit tests for known interfaces.
- Formatting code or modifying visual layouts.

## Required context and inputs

Before initiating research:

1. **The Research Objective:** Specific technical question or hypothesis to prove or disprove.
2. **Current Project Environment:** Runtime versions, framework versions, operating system, and target deployment platform.
3. **Constraints:** Licensing requirements (MIT, Apache 2, GPL limits), package size limits, security boundaries, and operational budget.
4. **Authoritative Sources:** Official documentation portals, repository URLs, and release note locations.

## Operating procedure

Execute the following ten-step technical research workflow:

1. **Define the Research Question:** Formulate a precise, bounded technical question (_"Does Next.js 15 support WebSocket streaming in App Router Route Handlers natively?"_).
2. **Identify What Must Be Verified:** List concrete facts needed (API signatures, runtime support, breaking changes, performance bounds).
3. **Prioritize Official Documentation:** Consult Tier 1 primary sources (MDN, official language/framework docs) first.
4. **Search for Current and Relevant Sources:** Ensure consulted sources match the exact major version used in the project. Reject outdated blog posts from superseded versions.
5. **Compare Competing Approaches:** If evaluating alternatives, analyze at least two viable solutions across performance, maintainability, community health, and license terms.
6. **Verify Version Compatibility:** Inspect `package.json` peer dependencies, node engine ranges, and native binary prerequisites.
7. **Identify Conflicting Information:** If sources disagree, cross-reference against official GitHub issue trackers, commit logs, or run a local reproduction script to verify ground truth.
8. **Distinguish Facts from Assumptions:** Label statements as verified facts, plausible inferences, or unverified assumptions.
9. **Summarize Findings:** Synthesize findings into clear technical prose with executive summary and detailed trade-off analysis.
10. **Provide Recommendations and Citations:** Conclude with an actionable recommendation, citing exact URLs, documentation chapters, and dates checked.

## Technical and engineering standards

- **Source Priority Rule:** Strictly respect the source hierarchy: Official Docs > Official Repos & Release Notes > Standards & Security Advisories > Reputable Tech Blogs > Community Forums.
- **Never Fabricate:** Never invent API parameters, method names, benchmarks, or compatibility claims. If an API is undocumented or unclear, label it explicitly as unverified.
- **Temporal Anchoring:** Always state the version and date of consulted evidence.
- **Reproducible Evidence:** Whenever evaluating an unclear edge-case behavior, write a small local proof-of-concept test script to establish definitive empirical evidence.

## Verification and evidence requirements

- Cite specific documentation sections, URLs, or official changelog entries for all factual assertions.
- Explicitly state which version of the software was inspected.
- State explicitly if any question could not be verified due to lack of public documentation or access.

## Expected outputs and deliverables

- **Technical Research Report:** Structured briefing containing executive summary, methodology, detailed findings, and recommendations.
- **Alternative Comparison Matrix:** Table comparing options across key technical dimensions.
- **Compatibility & Migration Advisory:** Clear warning of breaking changes, peer dependency constraints, or deprecation notices.
- **Source Citations:** Formatted list of authoritative references.

## Risks, boundaries and failure modes

- **Stale Information Hazard:** Applying advice written for an obsolete version that causes silent breakage or security flaws.
- **Marketing Bias:** Believing landing page claims of library authors without verifying open GitHub issues or real-world benchmarks.
- **Unverified AI Hallucination:** Generating fictional API signatures because they sound plausible.

## Progressive disclosure references

For source hierarchy standards, fact vs. assumption criteria, and verification templates:

- [Deep Research: Evidence Synthesis, Source Verification, and Fact-Checking Guide](./references/evidence-synthesis-guide.md)
Pliki w paczce (2)
  • SKILL.md 6 KB
  • references/evidence-synthesis-guide.md 3 KB

Wersje

RepozytoriumWynikLicencja
Ta strona · hmfareed/Frontier-Engineer90MIT
shepherdjerred/monorepo100GPL-3.0
3x-haust/ondra100MIT
glebis/claude-skills100MIT
drvoss/everything-copilot-cli100MIT
5dive-ai/skills100MIT
huayu-pro/skills100MIT
lovstudio/skills100MIT
bytedance/deer-flow100MIT

Migawka bazy z 13.11.2026 · wersja paczki 3c2760546fd3 (skrót SHA-256 całej paczki)

Pobrania w ostatnich 30 dniach: 0 · ranking

Zgłoś problem z tym skillem: szkodliwa treść, błędny opis albo prośba autora o usunięcie.

Podobne skille

  • 100technical-research
    Weryfikuje twierdzenia techniczne, porównuje opcje i bada nieznane biblioteki, API i narzędzia, z cytowanymi dowodami i oceną pewności.
  • 100research
    Prowadzi pogłębiony research rozwiązań technicznych: przeszukuje sieć, analizuje repozytoria GitHub i zbiera dowody o opcjach wdrożenia lub technologiach.
  • 100looking-up-docs
    Znajduje dokładną dokumentację bibliotek, API i frameworków w zgodnej wersji jednym przepływem wyszukiwania: sygnatury API, klucze konfiguracji, składnia i przykłady, także przez Context7.
  • 90research
    Bada wskazaną bibliotekę, framework, SDK, API lub usługę albo szuka przewodnika migracji, wysyłając agenta zwiadowcę.
  • 95ark-research
    Bada rozwiązania techniczne przez przeszukiwanie sieci, analizę repozytoriów GitHub i zbieranie dowodów, przy ocenie opcji implementacji i technologii.
  • 100research
    Prowadzi pogłębiony research techniczny z oceną wielu opcji przy wyborze technologii, decyzjach architektonicznych, porównaniu bibliotek oraz analizie wydajności.

Uwagi i zgłoszenia

Wiadomość trafi do Bartosza Sokolińskiego. Adres e-mail jest potrzebny tylko do odpowiedzi. Szczegóły w polityce prywatności.