AgentSkillsCN

gsd-researcher

在规划前开展技术调研，深入探究各类技术、生态系统与实现模式。在制定计划之前，先进行充分的研究与探索。

SKILL.md
--- frontmatter
name: gsd-researcher
description: Research methodology for investigating technologies, ecosystems, and implementation patterns. Invoke when researching before planning.

GSD Researcher

Systematic research methodology for discovering standard stacks, architecture patterns, and pitfalls.

Core Philosophy

Claude's training data is 6-18 months stale. Treat pre-existing knowledge as hypothesis, not fact.

  • •Verify before asserting
  • •Date your knowledge
  • •Prefer current sources
  • •Flag uncertainty

Research Modes

Mode 1: Ecosystem

When: "What tools exist for X?"

  • •What libraries/frameworks exist
  • •What's the standard stack
  • •What's SOTA vs deprecated

Mode 2: Feasibility

When: "Can we do X?"

  • •Is it technically achievable
  • •What constraints exist
  • •What's the effort/complexity

Mode 3: Implementation

When: "How do we implement X?"

  • •Specific patterns
  • •Code examples
  • •Configuration
  • •Common pitfalls

Mode 4: Comparison

When: "Compare A vs B"

  • •Feature comparison
  • •Performance comparison
  • •Clear recommendation

Source Hierarchy

LevelSourcesTrust
HIGHOfficial docs, releasesState as fact
MEDIUMMultiple sources agreeState with attribution
LOWSingle source, unverifiedFlag for validation

Priority order:

  1. •Official documentation
  2. •Official GitHub (README, releases)
  3. •Web search (verified)
  4. •Web search (unverified) — mark LOW

Research Process

Step 1: Identify Domains

  • •Core technology
  • •Ecosystem/stack
  • •Patterns
  • •Pitfalls
  • •SOTA check

Step 2: Execute Research

For each domain:

  1. •Check official docs first
  2. •Search for current best practices (include year)
  3. •Cross-verify findings
  4. •Document confidence levels

Step 3: Quality Check

  • • All items investigated
  • • Negative claims verified
  • • Multiple sources for critical claims
  • • URLs provided
  • • Publication dates checked
  • • Confidence levels assigned

Output: RESEARCH.md

markdown
# Phase [X]: [Name] - Research

**Domain:** [technology]
**Confidence:** [HIGH/MEDIUM/LOW]

## Summary

[What was researched, standard approach, key recommendations]

**Primary recommendation:** [one-liner]

## Standard Stack

| Library | Version | Purpose | Why Standard |
|---------|---------|---------|--------------|
| [name] | [ver] | [purpose] | [why] |

## Architecture Patterns

### Recommended Structure

src/ ├── [folder]/ # [purpose] └── [folder]/ # [purpose]

code

### Pattern: [Name]
**What:** [description]
**When:** [conditions]
**Example:** [code]

## Don't Hand-Roll

| Problem | Don't Build | Use Instead |
|---------|-------------|-------------|
| [problem] | [custom] | [library] |

## Common Pitfalls

### Pitfall: [Name]
**What goes wrong:** [description]
**How to avoid:** [prevention]

## Sources

- [URL] — [what, confidence]

Verification Protocol

Red Flags

  • •Every investigation succeeds perfectly (unrealistic)
  • •All findings equally certain (no confidence levels)
  • •"According to docs" without URL
  • •"X cannot do Y" without citation

Checklist

  • • All enumerated items investigated
  • • Negative claims verified with official docs
  • • Multiple sources for critical claims
  • • URLs provided for authoritative sources
  • • Publication dates checked
  • • Confidence levels assigned honestly
  • • "What might I have missed?" review