Search everything...

Stats

Actions

Available In

research-kit

Name: research-kit
Author: shine102

By shine102

AI co-worker and second brain for security engineers -- helps solve real tasks while building deep understanding through on-demand learning

npx claudepluginhub shine102/research-kit --plugin research-kit

Popularity

Stars

Med: 0·Avg: 285

Installs

Med: 0·Avg: 1

What's Inside

Agents3

collaborator

/collaborator

Use this agent as the default co-worker when the user starts a new security topic or continues existing work. Examples: <example>Context: User starts a new research topic. user: "/research-kit:start evaluate security for enterprise AI chatbot" assistant: "Let me dispatch the Collaborator to help break down this task." <commentary>New topic requires the Collaborator agent to analyze scope, check user knowledge, and plan the approach.</commentary></example> <example>Context: User is working through a topic and needs to produce a deliverable. user: "Create a threat model for this architecture" assistant: "The Collaborator will draft this deliverable based on our research so far." <commentary>Deliverable creation is core Collaborator responsibility.</commentary></example>

mentor

/mentor

Use this agent when the user needs to understand a security concept -- either detected as a knowledge gap by Collaborator or explicitly requested by the user. Examples: <example>Context: Collaborator detects user doesn't understand prompt injection at scale. user: "I'm not sure how prompt injection defense works at scale" assistant: "Let me bring in the Mentor to walk through this concept with you." <commentary>Knowledge gap detected during task work -- Mentor teaches the concept then hands back.</commentary></example> <example>Context: User directly wants to learn. user: "/research-kit:learn STRIDE threat modeling" assistant: "Dispatching the Mentor to teach STRIDE interactively." <commentary>User explicitly requests learning on a topic.</commentary></example>

researcher

/researcher

Use this agent for deep research that requires searching multiple sources, synthesizing findings, and producing structured research notes. Examples: <example>Context: Collaborator needs comprehensive research on a new topic. user: "I need to understand the current state of enterprise AI security frameworks" assistant: "Dispatching the Researcher to do a deep dive across authoritative and community sources." <commentary>Multi-source research on a broad topic requires dedicated Researcher agent.</commentary></example> <example>Context: User directly requests research. user: "/research-kit:deep-research latest prompt injection defense techniques" assistant: "Dispatching the Researcher for deep research on this topic." <commentary>User explicitly requests research mode.</commentary></example>

Skills11

assess-knowledge

/assess-knowledge

Use when needing to check user's familiarity with a topic before diving in, or after teaching to verify understanding -- reads and updates the knowledge state in .research-kit/profile/knowledge.md

create-exercise

/create-exercise

Use when the user needs hands-on practice after learning a concept -- generates exercises appropriate to the topic type (critical thinking, threat modeling, links to online labs)

curate-resources

/curate-resources

Use when needing to find and organize learning resources for a topic -- curates docs, labs, blogs, tools, and forum discussions matched to user's knowledge level

deep-research

/deep-research

Use when the user wants dedicated deep research on a security topic -- dispatches the Researcher agent for multi-source investigation with structured output

draft-deliverable

/draft-deliverable

Use when needing to produce a work output document -- threat models, playbooks, checklists, architecture docs, reports, and other security deliverables

Stats

Version0.1.0

Stars0

MaintenanceExcellent

LicenseMIT

Last CommitApr 12, 2026

AddedApr 12, 2026

Actions

View on GitHub View README Plugin Marketplace JSON

Own this plugin?

Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).

Available In

research-kit

Safety Signals

Caution

Uses power tools

Uses Bash, Write, or Edit tools

README

Research Kit

AI co-worker and second brain plugin for Claude Code. Helps security engineers solve real work tasks while building deep understanding through on-demand learning.

Install

From GitHub

# Add the marketplace
/plugin marketplace add shine102/research-kit

# Install the plugin
/plugin install research-kit@research-kit

For Development

# Run Claude Code with the plugin loaded from local directory
claude --plugin-dir /path/to/research-kit

Reload After Changes

/reload-plugins

Usage

/research-kit:start "evaluate security for enterprise AI chatbot"
/research-kit:start continue
/research-kit:learn "prompt injection defense"
/research-kit:deep-research "latest AI security frameworks 2026"

Entry Points

Command	Description
`/research-kit:start "topic"`	Start a new topic -- Collaborator agent breaks down the task
`/research-kit:start continue`	Resume most recent topic
`/research-kit:start list`	List all topics
`/research-kit:learn "concept"`	Learn a specific concept -- dispatches Mentor agent
`/research-kit:deep-research "query"`	Deep research -- dispatches Researcher agent

Agents

Collaborator -- default co-worker, breaks down tasks, detects knowledge gaps, produces deliverables
Mentor -- teaches concepts interactively when knowledge gaps detected, challenges thinking
Researcher -- deep multi-source research with trusted source priority

Skills

Skill	Description
`start`	Main entry point and agent orchestrator
`learn`	Direct access to Mentor agent
`deep-research`	Direct access to Researcher agent
`manage-workspace`	Workspace CRUD operations
`assess-knowledge`	Check and update user knowledge state
`web-search`	Search with trusted source weighting
`summarize`	Structured content summaries
`curate-resources`	Find and annotate resources by level
`teach`	Explain concepts interactively
`create-exercise`	Generate exercises by topic type
`draft-deliverable`	Produce work documents (threat models, playbooks, checklists, reports)

How It Works

You give a task (e.g., "evaluate security for our AI chatbot")
Collaborator agent breaks it down, asks clarifying questions
When a knowledge gap is detected, Mentor agent teaches the concept interactively
When deep research is needed, Researcher agent searches trusted sources
All progress, knowledge state, and deliverables are saved to .research-kit/ workspace

The plugin tracks your knowledge state (solid / learning / gap) and adapts its behavior accordingly -- like a second brain that knows what you know.

File Structure

research-kit/
  .claude-plugin/
    plugin.json                      # Plugin metadata
    marketplace.json                 # Marketplace config for GitHub install
  agents/
    collaborator.md                  # Default co-worker agent
    mentor.md                        # Teaching agent
    researcher.md                    # Deep research agent
  skills/
    start/SKILL.md                   # Main entry point
    learn/SKILL.md                   # Direct Mentor access
    deep-research/SKILL.md           # Direct Researcher access
    manage-workspace/SKILL.md        # Workspace operations
    assess-knowledge/SKILL.md        # Knowledge state management
    web-search/SKILL.md              # Trusted source search
    summarize/SKILL.md               # Content summarization
    curate-resources/SKILL.md        # Resource curation
    teach/SKILL.md                   # Interactive teaching
    create-exercise/SKILL.md         # Exercise generation
    draft-deliverable/SKILL.md       # Document drafting
  templates/
    topic-context.md
    roadmap.md
    progress.md
    sources.md
    settings.md

Workspace (created at runtime)

When you start your first topic, the plugin creates .research-kit/ in your working directory:

.research-kit/
  config/
    sources.md                       # Trusted sources registry (editable)
    settings.md                      # Plugin preferences
  profile/
    knowledge.md                     # Your knowledge state (solid/learning/gap)
    preferences.md                   # Communication preferences
    history.md                       # Past topics
  topics/
    <topic-slug>/
      context.md                     # Task brief
      roadmap.md                     # Sub-topic breakdown
      progress.md                    # Progress tracking
      sources.md                     # Topic-specific sources
      notes/                         # Research notes
      exercises/                     # Practice materials
      deliverables/                  # Output docs

License

MIT

research-kit

Popularity

What's Inside

Confidence

README

Research Kit

Install

From GitHub

For Development

Reload After Changes

Usage

Entry Points

Agents

Skills

How It Works

File Structure

Workspace (created at runtime)

License

Similar Plugins

security-awareness

faos-security-engineer

cybersec-toolkit

grimoire

cybersecurity-skills

security-toolkit

Research Kit

Install

From GitHub

For Development

Reload After Changes

Usage

Entry Points

Agents

Skills

How It Works

File Structure

Workspace (created at runtime)

License

Popularity

Health & Quality

Similar Plugins

security-awareness

faos-security-engineer

cybersec-toolkit

grimoire

cybersecurity-skills

security-toolkit