New: 98,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

Research Then Implement

by UniClipboard in UniClipboard/UniClipboard

Agent Loop for tasks that need research before implementation: check existing research first, run structured investigation (bb-browser/context7/code study), produce a Brief, then implement against…

AGPL-3.0Auto-check passedAgent Workflows

Install Research Then Implement

skills CLI
$ npx skills add UniClipboard/UniClipboard --skill research-then-implement -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install UniClipboard/UniClipboard research-then-implement --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/UniClipboard/UniClipboard.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/research-then-implement .claude/skills/research-then-implement && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
research-then-implement
GitHub stars
1.9k
Token cost
~2.4k tokens
SKILL.md length
630 words
Files
1
Skills in repo
24
Repo updated
First seen
Licence
AGPL-3.0

At a glance

Agent Loop for tasks that need research before implementation: check existing research first, run structured investigation (bb-browser/context7/code study), produce a Brief, then implement against…

  • Works in 4 steps: Check for existing research → Structured Investigation → Produce the Brief → …
  • Tasks that involve Autonomous loops
  • SKILL.md covers Purpose, When to trigger, When NOT to use and Directory convention, plus 7 more sections
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Research Then Implement is an agent skill from UniClipboard/UniClipboard. Agent Loop for tasks that need research before implementation: check existing research first, run structured investigation (bb-browser/context7/code study), produce a Brief, then implement against the Brief. Prevents repeated research on already-studied topics and research-to-implementation drift.

Its SKILL.md is about 2.4k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Agent Workflows, covering Autonomous loops. The repository describes itself as: Real-time clipboard sync across all your devices — local-first, peer-to-peer, and end-to-end encrypted. No account. No cloud dependency. No central server. The licence is AGPL-3.0.

When your agent uses it

  • Tasks that involve Autonomous loops

Example prompts

  • “/research-then-implement”

Workflow steps

4 steps, taken from the step headings in SKILL.md.

  1. Check for existing research
  2. Structured Investigation
  3. Produce the Brief
  4. Implement against the Brief

What it can do on your machine

Read from SKILL.md and the folder at commit 101ffb3. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md (its code samples are bash and markdown).

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Research Then Implement loads about 2.4k tokens when it runs. Until then it costs about 81 tokens; SKILL.md has 630 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~81
When it runs · the whole SKILL.md, loaded when a task matches
~2.4k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from UniClipboard/UniClipboard at commit 101ffb3, republished under its AGPL-3.0 licence (© UniClipboard). 630 words, ~2,437 tokens.

Download SKILL.mdSave it as .claude/skills/research-then-implement/SKILL.md (or your agent's skills folder).
name
research-then-implement
description
Agent Loop for tasks that need research before implementation: check existing research first, run structured investigation (bb-browser/context7/code study), produce a Brief, then implement against the Brief. Prevents repeated research on already-studied topics and research-to-implementation drift.

research-then-implement

Purpose

An Agent Loop for tasks where the right approach isn't obvious and requires investigation before coding. It prevents two recurring problems:

  1. Repeated research: Re-investigating topics that were already studied in prior sessions (e.g., researching Expo UI components twice because the first research wasn't captured)
  2. Research-implementation drift: Research conclusions are in the agent's context but not in a durable artifact, so the implementation silently diverges from what was learned

Before:

text
Session N: [researches via bb-browser, learns best practices] → implements → session ends
Session N+1: [re-researches the same topic from scratch because findings weren't saved]

After:

text
Session N: $research-then-implement → checks .planning/research/ → no prior research →
           researches → writes BRIEF.md → implements against BRIEF → done
Session N+1: $research-then-implement → checks .planning/research/ → finds BRIEF.md →
           skips research → implements directly

When to trigger

  • $research-then-implement or /rti — start the full loop
  • $research-then-implement research-only — just do the research, save Brief, don't implement
  • User says "先研究一下再做", "research this first", "我不确定怎么做"
  • Tasks involving unfamiliar frameworks, APIs, or design patterns
  • Cross-platform porting (iOS → Android, native → Expo)
  • Evaluating multiple approaches before choosing one

When NOT to use

  • The implementation approach is already clear → just do it
  • Bug fixes with known root cause → use systematic-debugging or error-diagnose-fix
  • Pure research without implementation → use deep-research skill instead

Directory convention

text
.planning/research/
  <topic-slug>/
    BRIEF.md          ← the deliverable: findings + recommendation + constraints
    sources.md        ← raw notes, URLs, code snippets from investigation

This convention already partially exists in the project (.planning/research/ has prior work). The skill formalizes it.

Phase 1 — Check for existing research

Before doing any new research, check if the topic was already studied:

bash
# Check .planning/research/ for matching topics
ls .planning/research/ 2>/dev/null

# Check memory for relevant entries
grep -i "<topic keywords>" \
  ${CODEX_HOME:-$HOME/.codex}/memories/MEMORY.md 2>/dev/null

# Check if there's a planning doc that covers this
find .planning -name "*.md" -newer .planning/research 2>/dev/null | \
  xargs grep -li "<topic>" 2>/dev/null

If existing research is found:

text
📚 Found existing research on this topic:
  .planning/research/expo-ui-migration/BRIEF.md (2026-06-21)

  Summary: Expo UI (@expo/ui) provides native iOS components...
  Recommendation: Use @expo/ui for iOS, custom components for Android
  
  A) Use this research and proceed to implementation
  B) Research is stale — redo from scratch
  C) Supplement — do additional research on specific gaps

Phase 2 — Structured Investigation

If no existing research, or user wants fresh research, run the investigation.

2a — Define the research question

Clarify what needs to be learned:

text
Research question: How to implement a native-feel bottom sheet in Expo
                   that matches iOS UISheetPresentationController behavior?

Sub-questions:
  1. Does @expo/ui provide a native BottomSheet component?
  2. What's the standard pattern for iOS Sheet in React Native/Expo?
  3. How does the iOS 26 Liquid Glass design affect this?
  4. What are the platform-specific considerations (iOS vs Android)?
2b — Multi-source investigation

Use the appropriate tool for each source type:

Official documentation (context7):

text
mcp__context7__resolve-library-id("expo-ui")
mcp__context7__query-docs(id, "BottomSheet sheet presentation")

Design trends and community practices (bb-browser):

bash
bb-browser site google/search "expo bottom sheet native iOS 2026 best practice"
bb-browser open <relevant-url>
bb-browser eval "document.querySelector('article')?.innerText?.substring(0, 5000)"
bb-browser close

Reference implementations (code study):

bash
# If porting from an existing implementation:
find /Users/mark/MyProjects/iOSApp/UniClipboard -name "*.swift" | \
  xargs grep -l "sheet\|presentation" 2>/dev/null

# Read the reference implementation
Read <file>

Package ecosystem:

bash
# Check what's available
bb-browser site google/search "npm expo bottom sheet native 2026"
2c — Record raw findings

Write sources and raw notes to sources.md:

markdown
# Sources: <topic>

## Official docs
- expo-ui BottomSheet: [url] — supports detents, grabber, native iOS sheet
- ...

## Community
- Blog post: [url] — recommends X over Y because...
- GitHub issue: [url] — known limitation with...

## Reference implementation
- iOS app uses UISheetPresentationController with custom detents
- File: /Users/mark/.../ServerSwitcherView.swift:42-80
- Key pattern: .presentationDetents([.medium, .large])

## Raw code snippets
...

Phase 3 — Produce the Brief

The Brief is the key deliverable — a concise, actionable document that locks the research findings.

BRIEF.md template
markdown
# Brief: <topic>

**Date:** <date>
**Status:** Draft | Reviewed | Locked
**Research question:** <the original question>

## Recommendation

<1-3 sentences: what to do and why>

## Key findings

1. <finding 1 — with source reference>
2. <finding 2>
3. <finding 3>

## Approach

<The specific implementation approach chosen>

### What to use
- <library/API/pattern>: <why>

### What NOT to use
- <rejected alternative>: <why not>

## Constraints

- <constraint from docs/API limitations>
- <constraint from project architecture>
- <platform-specific constraint>

## Implementation checklist

- [ ] <step 1>
- [ ] <step 2>
- [ ] <step 3>

## Cross-platform considerations

| Aspect | iOS | Android |
|--------|-----|---------|
| Component | ... | ... |
| Behavior | ... | ... |

## Open questions

- <anything not resolved by research>
Brief quality gate

Before proceeding to implementation, verify:

  1. The recommendation is specific enough to implement (not "use the best library")
  2. At least one alternative was considered and rejected with reasons
  3. Constraints are concrete (not "be careful about performance")
  4. The checklist has actionable items

If the user passed research-only, stop here.

Phase 4 — Implement against the Brief

4a — Load the Brief as constraints

The Brief is now the specification. Implementation must:

  • Follow the recommended approach
  • Respect all listed constraints
  • Use the specified libraries/APIs (not alternatives)
  • Check off items from the implementation checklist
Show full SKILL.md (256 more words)Show less
4b — Implement

Standard implementation, but with one key rule:

If during implementation you discover the Brief's recommendation doesn't work (API doesn't exist, behavior differs from docs, breaking change):

  1. STOP — don't silently work around it
  2. Update the Brief with the new finding
  3. Tell the user: "The Brief said X, but in practice Y. Updated the Brief. Proceeding with Z instead."
4c — Mark Brief as complete

After implementation:

markdown
**Status:** Locked
**Implemented:** 2026-06-21
**Branch:** feature/bottom-sheet-native

Update the implementation checklist with completion marks.

Cross-platform porting (special case)

When the task is porting from one platform to another (iOS → Expo, Swift → React Native):

Transfer Spec

Before implementation, create a Transfer Spec that maps source → target:

markdown
## Transfer Spec: iOS → Expo BottomSheet

| iOS (source) | Expo (target) | Notes |
|-------------|---------------|-------|
| UISheetPresentationController | @expo/ui Sheet | Native on iOS |
| .presentationDetents([.medium]) | snapPoints={['50%']} | Different API |
| .prefersGrabberVisible(true) | enableGrabber={true} | Same concept |
| UIAction in menu | Context menu from @expo/ui | Not 1:1 |

### Behavioral parity checklist
- [ ] Half-screen snap point
- [ ] Full-screen expansion
- [ ] Drag-to-dismiss
- [ ] Grabber indicator
- [ ] Background dimming

This Transfer Spec is reusable — if the same port needs refinement later, the mapping is already done.

Interaction with other skills

SkillRelationship
deep-researchFor pure research questions without implementation intent
bb-browserTool used in Phase 2 for web research
context7Tool used in Phase 2 for official docs
$wrapShould persist the Brief path in context_refs
$continue-taskShould read the Brief on resume
codebase-designMay inform the Brief's constraints section

Anti-patterns

  • Researching and implementing in the same mental breath (research, then implement — don't interleave)
  • Re-researching a topic that has a valid Brief in .planning/research/
  • Writing a Brief so vague it doesn't constrain implementation ("use best practices")
  • Silently deviating from the Brief during implementation without updating it
  • Spending 3 rounds researching when the topic is well-documented (use context7 first)
  • Researching without producing a Brief (findings evaporate with the session)
  • Porting code line-by-line instead of understanding the design intent first

© UniClipboard, AGPL-3.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in .agents/skills/research-then-implement of UniClipboard/UniClipboard.

Open the folder on GitHubat commit 101ffb3

Compare with similar skills

Research Then Implement next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Research Then Implement compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Research Then Implement this skillUniClipboard/UniClipboard1.9k—~2.4kAutomated safety check: PassAGPL-3.0
Autoresearch Iteration Loopuditgoenka/autoresearch6.5k1 repos~2kAutomated safety check: PassMIT
PUA Looptanweai/pua20k1 repos~1.1kAutomated safety check: PassMIT
Install Loop Engineeringcobusgreyling/loop-engineering11k1 repos~648Automated safety check: PassMIT
LoopX Self Repairloopx-project/loopx6.2k—~2.2kAutomated safety check: PassApache-2.0
PRP LoopWirasm/prp2.3k—~894Automated safety check: PassMIT

Similar skills

  • Autoresearch Iteration Loop

    uditgoenka/autoresearch

    Runs an autonomous modify, verify, keep-or-discard loop against any metric, with subcommands for planning, debugging, fixing, security audits, shipping and more.

    6.5k GitHub starsUsed in 1 repo~2k tokens
    Agent WorkflowsAuto-check passed

  • PUA Loop

    tanweai/pua

    Runs an unattended iterate-until-verified loop in which a user-set verify command, not the agent's own claim, decides when the task is finished.

    20k GitHub starsUsed in 1 repo~1.1k tokens
    Agent WorkflowsAuto-check passed

  • Install Loop Engineering

    cobusgreyling/loop-engineering

    Installs Loop Engineering into a project through the single @cobusgreyling/loop CLI, scaffolding a report-only loop and a readiness score.

    11k GitHub starsUsed in 1 repo~648 tokens
    Agent WorkflowsAuto-check passed

  • LoopX Self Repair

    loopx-project/loopx

    Diagnoses surprising LoopX behavior, such as stale recommendations or tiny progress, assigns it to the responsible layer and repairs it at the lowest durable level.

    6.2k GitHub stars~2.2k tokensUpdated today
    Agent WorkflowsAuto-check passed

  • PRP Loop

    Wirasm/prp

    Runs the plan, implement and review pipeline detached in fresh headless sessions, looping review and fix until the pull request is clean.

    2.3k GitHub stars~894 tokensUpdated 6 days ago
    Agent WorkflowsAuto-check passed

  • Self-Improve Evolutionary Loop

    Yeachan-Heo/oh-my-claudecode

    Runs an autonomous improvement loop on a repository: agents propose and execute plans, a tournament picks the winner by benchmark, and each round is recorded and plotted.

    40k GitHub stars~5.3k tokensUpdated today
    Agent WorkflowsAuto-check: warnings

More from UniClipboard/UniClipboard

All 24 skills in this repo

  • Beui

    UniClipboard/UniClipboard

    Pick and install beUI (@beui) animated React components from the shadcn registry.

    1.9k GitHub stars~1.8k tokensUpdated today
    Auto-check passed

  • Create PR

    UniClipboard/UniClipboard

    Push the current branch and open a GitHub pull request against main.

    1.9k GitHub stars~4.6k tokensUpdated today
    Auto-check passed

  • Design Audit

    UniClipboard/UniClipboard

    定期审计代码库的工程设计问题（高心智复杂度、单一真相源被破坏、catch-all 胖接口、死代码、散落魔法字面量、泄漏抽象、资源生命周期靠环形缓冲）与可优化点，范围限定为自上次审计以来的 git churn，每条发现都落到 file:line 并对照本项目自己的 VISION.md / 各级 AGENTS.md / memory…

    1.9k GitHub stars~554 tokensUpdated today
    Auto-check passed

  • Dual Side Debug

    UniClipboard/UniClipboard

    Inspect uniclipboard logs from BOTH the macOS host and the mounted Windows peer when debugging cross-platform sync, pairing, transfer, or daemon issues.

    1.9k GitHub stars~2.6k tokensUpdated today
    Auto-check passed

  • E2E Test Thinker

    UniClipboard/UniClipboard

    Analyze the current branch's diff against main and determine which changes are testable via CLI-based end-to-end tests.

    1.9k GitHub stars~1.7k tokensUpdated today
    Auto-check passed

  • iOS Log Diagnose

    UniClipboard/UniClipboard

    Drive the UniClipboard iOS app in a simulator and read its OSLog yourself to diagnose a mobile-sync bug, instead of asking the user to paste logs.

    1.9k GitHub stars~1.1k tokensUpdated today
    Auto-check passed

Categories

Questions about Research Then Implement

What does Research Then Implement do?

Agent Loop for tasks that need research before implementation: check existing research first, run structured investigation (bb-browser/context7/code study), produce a Brief, then implement against…. Research Then Implement is an agent skill from UniClipboard/UniClipboard. Agent Loop for tasks that need research before implementation: check existing research first, run structured investigation (bb-browser/context7/code study), produce a Brief, then implement against the Brief.

When should I use Research Then Implement?

Research Then Implement fits situations like: tasks that involve Autonomous loops.

How do I install Research Then Implement in Claude Code?

Run `npx skills add UniClipboard/UniClipboard --skill research-then-implement -a claude-code`. Or copy the skill folder (.agents/skills/research-then-implement in UniClipboard/UniClipboard) into .claude/skills/research-then-implement in your project. Claude Code loads it when a task matches its description.

How do I install Research Then Implement in Codex?

Run `npx skills add UniClipboard/UniClipboard --skill research-then-implement -a codex`. Or copy the skill folder (.agents/skills/research-then-implement in UniClipboard/UniClipboard) into .agents/skills/research-then-implement in your project. Codex loads it when a task matches its description.

Can I use Research Then Implement in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add UniClipboard/UniClipboard --skill research-then-implement -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/research-then-implement, .gemini/skills/research-then-implement, .github/skills/research-then-implement and .opencode/skills/research-then-implement in your project.

What does Research Then Implement need to run?

SKILL.md names no scripts, command-line tools or credentials: Research Then Implement is instructions for the agent only.

Does Research Then Implement access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Research Then Implement safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Research Then Implement use?

Research Then Implement is published under the AGPL-3.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Research Then Implement use?

About 2.4k tokens (SKILL.md is roughly 9.7k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Research Then Implement?

Skills that share tags, products or a category with Research Then Implement: Autoresearch Iteration Loop (uditgoenka/autoresearch, 6.5k stars), PUA Loop (tanweai/pua, 20k stars), Install Loop Engineering (cobusgreyling/loop-engineering, 11k stars) and LoopX Self Repair (loopx-project/loopx, 6.2k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Research Then Implement?

UniClipboard (a GitHub organization) maintains it in UniClipboard/UniClipboard, which has 1,856 GitHub stars. The repository holds 24 skills in this directory. The repository was last updated on October 8, 2026.

Source: UniClipboard/UniClipboard on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.