playbooks

home / skills / yeachan-heo / oh-my-claudecode / external-context

external-context skill

/skills/external-context

This skill orchestrates parallel external context searches by routing facets to document-specialist agents and synthesizing results for actionable insights.

npx playbooks add skill yeachan-heo/oh-my-claudecode --skill external-context

Review the files below or copy the command above to add this skill to your agents.

Files (1)
SKILL.md
2.4 KB
---
name: external-context
description: Invoke parallel document-specialist agents for external web searches and documentation lookup
argument-hint: <search query or topic>
---

# External Context Skill

Invoke parallel document-specialist agents to search the web for external documentation, references, and context.

## Overview

External Context decomposes a query into parallel web search facets, each handled by an independent document-specialist agent:

1. **Decomposition** - Break query into 2-5 independent search facets
2. **Parallel Search** - Spawn document-specialist agents for each facet
3. **Synthesis** - Aggregate findings into structured context

## Usage

```
/oh-my-claudecode:external-context <topic or question>
```

### Examples

```
/oh-my-claudecode:external-context What are the best practices for JWT token rotation in Node.js?
/oh-my-claudecode:external-context Compare Prisma vs Drizzle ORM for PostgreSQL
/oh-my-claudecode:external-context Latest React Server Components patterns and conventions
```

## Protocol

### Facet Decomposition

Given a query, decompose into 2-5 independent search facets:

```markdown
## Search Decomposition

**Query:** <original query>

### Facet 1: <facet-name>
- **Search focus:** What to search for
- **Sources:** Official docs, GitHub, blogs, etc.

### Facet 2: <facet-name>
...
```

### Parallel Agent Invocation

Fire independent facets in parallel via Task tool:

```
Task(subagent_type="oh-my-claudecode:document-specialist", model="sonnet", prompt="Search for: <facet 1 description>. Use WebSearch and WebFetch to find official documentation and examples. Cite all sources with URLs.")

Task(subagent_type="oh-my-claudecode:document-specialist", model="sonnet", prompt="Search for: <facet 2 description>. Use WebSearch and WebFetch to find official documentation and examples. Cite all sources with URLs.")
```

### Synthesis

After all agents complete, synthesize findings:

```markdown
## External Context: <query>

### Key Findings
1. **<finding>** - Source: [title](url)
2. **<finding>** - Source: [title](url)

### Detailed Results

#### Facet 1: <name>
<aggregated findings with citations>

#### Facet 2: <name>
<aggregated findings with citations>

### Sources
- [Source 1](url)
- [Source 2](url)
```

## Configuration

- Maximum 5 parallel document-specialist agents
- Each agent uses WebSearch and WebFetch tools
- No magic keyword trigger - explicit invocation only

Overview

This skill invokes parallel document-specialist agents to collect external web documentation and references for a given topic or question. It decomposes queries into focused search facets, runs independent web searches in parallel, and synthesizes results into structured, citeable context. The outcome is a concise, source-backed summary you can use inside a team workflow.

How this skill works

Given a user query, the skill splits it into 2–5 independent search facets that cover distinct aspects of the topic. It then spawns document-specialist agents in parallel, each using web search and fetch tools to gather official docs, examples, and blog posts. When all agents finish, the skill aggregates findings, extracts key takeaways with source citations, and returns a structured report with detailed facet-level results and a consolidated source list.

When to use it

  • Researching best practices or patterns with up-to-date web sources
  • Comparing libraries, frameworks, or tools with external documentation and benchmarks
  • Gathering examples and authoritative guides for implementation questions
  • Preparing technical summaries or references for team knowledge sharing
  • Validating claims by collecting primary source citations

Best practices

  • Be specific in the original query so facets target independent, non-overlapping areas
  • Limit scope to a few focused facets (2–5) to keep searches precise and reviewable
  • Require agents to cite all sources with URLs for traceability
  • Use the synthesized report as a starting point; inspect individual facet results before final decisions
  • Invoke explicitly — the skill runs only when called to avoid accidental web fetching

Example use cases

  • Ask for best practices and citations for token rotation in Node.js to support a security design doc
  • Compare two ORMs for PostgreSQL with links to benchmarks, docs, and migration guides
  • Collect implementation patterns and official guidance for React Server Components
  • Produce a short, sourced briefing on a third-party API before integration work
  • Gather regulatory or standards references when drafting compliance notes

FAQ

How many parallel agents will run for a single query?

Up to five document-specialist agents can run in parallel; typical decompositions use 2–5 facets.

What sources do agents use and are sources cited?

Agents use web search and web fetch tools to find official docs, GitHub, blog posts, and examples; every finding includes source citations and URLs.