research

ark-ai-agent

A modular marketplace of Claude Code plugins for planning, prompting, reflection, and productivity workflows. Install only what you need — each plugin is independent with its own commands, skills, and agents.

Forked from TACHES with added playbook system (ACE pattern) for cross-session knowledge accumulation.

Plugin Catalog

Plugin	What it does	Commands	Skills	Agents
consider	Mental model thinking frameworks	12	—	—
research	Structured research workflows	8	—	—
reflection	ACE playbook system	2	2	1
extensibility	Create/audit/heal Claude Code extensions	9	6	3
planning	Project planning + team orchestration	4	1	2
todos	Linear-integrated task management	3	—	—
tools	Docs, debugging, Ralph loops	3	3	—
prompting	Prompt engineering for delegation	3	—	—
expertise	Domain knowledge (iOS, macOS, n8n)	—	3	—
native-workflow	Native plan mode iterative workflow	3	—	1

Installation

Selective Install (Recommended)

From within Claude Code:

# Add the marketplace
/plugin marketplace add HelloWorldSungin/ark-ai-agent

# Install only plugins you need
/plugin install consider@ark-ai-agent
/plugin install planning@ark-ai-agent
/plugin install reflection@ark-ai-agent

Or from a local submodule:

/plugin marketplace add ./external/ark-ai-agent
/plugin install consider@ark-ai-agent

Use the interactive /plugin UI to browse all available plugins.

Migration from v2.x

v3.0.0 replaces the monolithic ark-ai-agent plugin with 9 independent plugins. To migrate:

1. Remove the old plugin: /plugin uninstall ark-ai-agent@ark-ai-agent
2. Install the plugins you need (see catalog above)
3. Commands that were ark-ai-agent:consider:first-principles are now consider:first-principles

---

consider

Apply mental frameworks to decisions and problems.

Command	Description
first-principles	Break down to fundamentals and rebuild
inversion	Solve backwards — what guarantees failure?
pareto	Apply 80/20 rule to focus on what matters
second-order	Think through consequences of consequences
5-whys	Drill to root cause
occams-razor	Find simplest explanation
one-thing	Identify highest-leverage action
swot	Map strengths, weaknesses, opportunities, threats
eisenhower-matrix	Prioritize by urgent/important
10-10-10	Evaluate across time horizons
opportunity-cost	Analyze what you give up
via-negativa	Improve by removing

research

Structured research workflows with web search and analysis.

Command	Description
competitive	Competitive analysis
deep-dive	Deep dive into a topic
feasibility	Feasibility study
history	Historical research
landscape	Landscape survey
open-source	Open source research
options	Options analysis
technical	Technical research

reflection

Accumulate reusable knowledge across sessions using the ACE (Agentic Context Engineering) pattern.

Component	Name	Description
Command	session-reflect	Analyze session and extract learnings
Command	playbook-update	Apply reflection to playbook files
Skill	session-reflect	End-of-session knowledge extraction
Skill	playbook-update	Curator-managed playbook updates
Agent	playbook-curator	ACE-style playbook curation with dedup

Workflow: End of session → /session-reflect → /playbook-update → knowledge persists to next session.

See playbook system docs for details.

extensibility

The complete Claude Code extension lifecycle: create, audit, and heal.