agentsclimarketplace

RECON

Skill jord0-cmd/jord0.skills/skills/RECON

Deep technical research that persists. Use when: exploring new technologies, investigating frameworks, comparative analysis of tools, understanding complex systems, or any research task that should be saved for future reference. Invokes Claude's built-in research agent and auto-saves results to a knowledge base for retrieval across sessions.From its SKILL.md

Install
npx -y skills add jord0-cmd/jord0.skills --skill RECON

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

4.3 KB, 926 tokens by cl100k_base, as published. Nobody here has run it

RECON

Deep research. Persistent results. Never research the same thing twice.


Usage

/recon [topic]                     - Research a topic in depth
/recon --quick [topic]             - Quick overview (5-10 minutes)
/recon --deep [topic]              - Comprehensive investigation (30+ minutes)
/recon --compare [A] vs [B]       - Comparative analysis

What This Is

RECON is a research skill that produces thorough technical reports and automatically saves them to a local knowledge base. Next time you (or Claude) need the same information, it's already there — no redundant web searches, no lost research.


The Research Process

1. Check Existing Knowledge

Before researching, RECON checks the knowledge base for existing reports on the topic. If a relevant report exists, it surfaces that first and asks if an update is needed.

2. Conduct Research

Uses Claude's built-in technical-research-agent to:

  • Decompose the topic into subtopics
  • Search multiple sources (web, documentation, repos)
  • Synthesize findings into a structured report
  • Identify code examples and implementation patterns

3. Save Results

Reports are saved as markdown files to the knowledge base directory:

knowledge/
  research/
    2024-01-15_react-server-components.md
    2024-01-20_rust-vs-go-backends.md
    2024-02-01_vector-databases-comparison.md
  index.json    # Master index for search

4. Index for Future Retrieval

The report is indexed with:

  • Title, topic, date
  • Key technologies mentioned
  • Summary for quick scanning
  • Tags for semantic search

AUTO-EXECUTE Protocol

When this skill is invoked:

  1. Parse the topic from the user's message
  2. Check knowledge base — search knowledge/index.json for existing research
  3. If found: Display the existing report summary, ask if an update is needed
  4. If not found: Launch the technical-research-agent (Task tool, subagent_type: technical-research-agent)
  5. Save the report to knowledge/research/YYYY-MM-DD_<topic-slug>.md
  6. Update the index with metadata
  7. Summarize findings to the user

Report Structure

Every RECON report follows this format:

# [Topic]

## Summary
One-paragraph executive summary.

## Key Findings
- Finding 1
- Finding 2
- Finding 3

## Deep Dive
Detailed analysis with code examples where relevant.

## Comparison (if applicable)
| Feature | Option A | Option B |
|---------|----------|----------|
| ...     | ...      | ...      |

## Recommendations
What to use and why, given the context.

## Sources
Links to documentation, articles, repos referenced.

Knowledge Base Setup

Create the following directory structure:

knowledge/
  research/      # Research reports (markdown)
  references/    # Quick reference docs
  index.json     # Master index

The index.json schema:

{
  "entries": [
    {
      "id": "k-001",
      "title": "React Server Components Deep Dive",
      "file": "research/2024-01-15_react-server-components.md",
      "date": "2024-01-15",
      "tags": ["react", "rsc", "frontend", "server-components"],
      "summary": "One-line summary for quick scanning"
    }
  ]
}

When to Use

  • Evaluating a new framework or library
  • Comparing architectural approaches
  • Understanding a technology before prototyping
  • Any research you don't want to repeat next week

Prerequisites

  • Claude Code with Task tool access (the technical-research-agent is a built-in Claude Code subagent — no external setup needed)
  • File system access for knowledge base storage
  • Web search access for current information (WebSearch tool must be available)

Note: The technical-research-agent is part of Claude Code itself, not an external service. If you have Claude Code, you have the agent.


Research once. Remember forever.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Gives 0 of the 12 instructions most context ai engineering skills give in 926 tokens

Counted across 1,193 of the 1,976 authors here whose files we hold, read 2026-08-07

  • Dispatch a fresh implementer subagent per taskin 48 of 1193, across 19 files
  • Dispatch a final code reviewer after all tasksin 33 of 1193, across 8 files
  • Provide full task text to the subagentin 30 of 1193, across 9 files
  • Review spec compliance before code qualityin 27 of 1193, across 10 files
  • Make the hook script executablein 26 of 1193, across 8 files
  • Re-snapshot after navigation or DOM changesin 25 of 1193, across 19 files
  • Read files before editing themin 22 of 1193, across 11 files
  • Answer subagent questions before proceedingin 22 of 1193, across 7 files
  • Mark task complete in TodoWrite after approvalin 22 of 1193, across 6 files
  • Merge hook into existing settingsin 21 of 1193, across 3 files
  • Ask if installation is global or projectin 20 of 1193, across 2 files
  • Copy the hook script to target locationin 20 of 1193, across 2 files

Said here and by no other author read

  • parse the topic from the message
  • search the knowledge base index for existing research
  • display existing reports before conducting new research
  • ask if an update is needed for existing reports
  • launch the technical research agent
  • save the report to the knowledge base

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 326,696. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.