search-first
Use before writing custom code or a research workflow when a maintained library, tool, skill, or proven pattern may already exist. Start with local reuse and targeted evidence; broaden to external research for new dependencies, new mechanisms, or uncertain external facts.
sidiangongyuan · 已有静态检查记录
安装与使用
运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。
npx skills add sidiangongyuan/codex-skills-library --skill search-first技能说明
name: search-first description: Use before writing custom code or a research workflow when a maintained library, tool, skill, or proven pattern may already exist. Start with local reuse and targeted evidence; broaden to external research for new dependencies, new mechanisms, or uncertain external facts. license: MIT
/search-first — Research Before You Code
Systematizes the "search for existing solutions before implementing" workflow.
Trigger
Use this skill when:
- Starting a new feature that likely has existing solutions
- Adding a dependency or integration
- The user asks "add X functionality" and you're about to write code
- Before creating a new utility, helper, or abstraction
Workflow
┌─────────────────────────────────────────────┐
│ 0. NEED + LOCAL REUSE GATE │
│ Inspect approved local implementations │
│ and results before external searching │
├─────────────────────────────────────────────┤
│ 1. RESEARCH DECISION │
│ Decide whether reuse, targeted lookup, │
│ or broader research is actually needed │
├─────────────────────────────────────────────┤
│ 2. TARGETED SEARCH (only when warranted) │
│ ┌──────────┐ ┌──────────┐ ┌──────────┐ │
│ │ Registry │ │ Tools / │ │ GitHub / │ │
│ │ / docs │ │ skills │ │ evidence │ │
│ └──────────┘ └──────────┘ └──────────┘ │
├─────────────────────────────────────────────┤
│ 3. EVALUATE │
│ Score candidates (functionality, maint, │
│ community, docs, license, deps) │
├─────────────────────────────────────────────┤
│ 4. DECIDE │
│ ┌─────────┐ ┌──────────┐ ┌─────────┐ │
│ │ Adopt │ │ Extend │ │ Build │ │
│ │ as-is │ │ /Wrap │ │ Custom │ │
│ └─────────┘ └──────────┘ └─────────┘ │
├─────────────────────────────────────────────┤
│ 5. IMPLEMENT │
│ Reuse approved solution or write the │
│ smallest justified custom code │
└─────────────────────────────────────────────┘
Decision Matrix
| Signal | Action |
|---|---|
| Approved local implementation or result | Reuse or repair — no broad fresh search required |
| Specific local gap or uncertainty | Targeted lookup — search only for that issue |
| New dependency, new mechanism, or uncertain external fact | Research — compare evidence before deciding |
| Exact match, well-maintained, MIT/Apache | Adopt — use directly when available or within authorized implementation scope |
| Partial match, good foundation | Extend — add a thin wrapper within authorized implementation scope |
| Multiple weak matches | Compose — combine 2-3 small packages |
| Nothing suitable found | Build — write custom, but informed by research |
Research Gate
Start with the repository and approved local catalog. Inspect existing modules, tests, lockfiles, configured tools or skills, prior results, and known patterns before searching externally. If an approved implementation or result already meets the need, reuse it. For a known defect or result gap, repair or fill it and use a targeted lookup only when a specific external fact or compatibility question remains; do not force a broad external search or launch the full researcher workflow for every fix.
Broader research is warranted when selecting a new dependency, introducing a new mechanism or architecture, or relying on external facts that may be uncertain or changed. Delegate only a bounded, independent search that would materially improve the decision. For search-only or advisory work, do not turn a recommendation into an automatic package install, configuration change, publication, message, or other external mutation. For an implementation request, ordinary project-scoped dependency installation and configuration are allowed within the existing task authority and budget; ask before global or system changes, new external effects, or scope or cost expansion.
How to Use
Step 0: Tool Availability Preflight
This is agent guidance, not an executable setup script. Check only the channels that are relevant to the task and project in front of you.
| Channel | Check | If missing |
|---|---|---|
| Repository search | rg --files and targeted rg queries | State that only visible files were inspected |
| Package registry | npm --version, python -m pip --version, or project package manager | Use web/docs search and avoid claiming registry coverage |
| GitHub CLI | gh auth status | Use public web or local git history only |
| MCP/docs tools | Available tool list or local MCP config | Fall back to official docs/web search |
| Skill catalog | Inspect the active harness's available skills or configured skill root | Say no local skill catalog was available |
Academic Literature Path
When the task needs research grounding, method positioning, related work,
citation support, or top-venue precedent, use $research-evidence instead of
ad hoc web search. Prefer CVPR, ICCV, ECCV, ICLR, NeurIPS/NIPS, and ICML for
CV/ML work; include CoRL, ICRA, IROS, AAAI, IJCAI, T-ITS, and RA-L only when
the user topic justifies autonomous-driving, robotics, or collaborative
perception coverage.
Use the evidence result to decide Adopt, Extend, Compose, or Build only
when the Research Gate calls for external evidence.
Do not claim literature coverage when the research-evidence tool or source
channel was unavailable.
Quick Mode (inline)
Before writing a utility or adding functionality, mentally run through:
- Is there an approved implementation or result locally? → Inspect relevant modules, tests, configs, and outputs first; reuse or repair it.
- Is there a specific gap or uncertainty? → Run a targeted lookup only if it would change the decision.
- Is a new dependency or mechanism, or an uncertain external fact involved? →