Skip to content

Concepts Overview

High-level understanding of Tachikoma's architecture and design philosophy.

What is Tachikoma?

Tachikoma is a research-backed AI agent orchestration system that:

  • Classifies intent — Understands what you want
  • Routes appropriately — Selects optimal execution strategy
  • Executes with quality — Uses PAUL methodology and CARL quality gates
  • Reflects on outcomes — Freedom to question and revise

Core Philosophy

Structure at the Start, Freedom at the End

Tachikoma provides structure during execution (phases 1-4) and freedom at the end (phase 5):

1. Classify Intent ← Structured
2. Load Context   ← Structured
3. Load Skill     ← Structured
4. Execute        ← Structured
5. Reflect        ← Freedom to question

Why:

  • Structure ensures quality and consistency
  • Freedom allows for creativity and improvement
  • Balance between automation and agency

Research-Backed Design

Every major feature is grounded in peer-reviewed research:

FeatureResearchSource
Intent RoutingCost-Aware RoutingarXiv:2601.02663
Context ManagementPosition BiasACL 2024
Verification LoopsGenerator-Verifier-ReviserarXiv:2602.10177
Model-Aware EditingModel HarnessCan Bouluk, Feb 2026
Large ContextRLMarXiv:2512.24601
Skill CompositionModularityarXiv:2602.03279

Learn more about the research →

System Architecture

Five-Phase Execution

┌─────────────────────────────────────────────────────────────┐
│                    User Request                             │
└──────────────────────────┬──────────────────────────────────┘

┌─────────────────────────────────────────────────────────────┐
│  1. CLASSIFY INTENT                                         │
│  - Extract intent keywords                                  │
│  - Match against routes                                     │
│  - Evaluate confidence                                      │
│  - Select execution strategy                                │
└──────────────────────────┬──────────────────────────────────┘

┌─────────────────────────────────────────────────────────────┐
│  2. LOAD CONTEXT                                            │
│  - Load system prompt                                       │
│  - Load skill instructions                                  │
│  - Load AGENTS.md (project rules)                           │
│  - Load context module (if applicable)                      │
│  - Load relevant files                                      │
└──────────────────────────┬──────────────────────────────────┘

┌─────────────────────────────────────────────────────────────┐
│  3. LOAD SKILL                                              │
│  - Discover skill from .opencode/skills/                    │
│  - Load SKILL.md instructions                               │
│  - Select model-aware edit format                           │
│  - Prepare tools                                            │
└──────────────────────────┬──────────────────────────────────┘

┌─────────────────────────────────────────────────────────────┐
│  4. EXECUTE (PAUL)                                          │
│  - PLAN: Define objectives, acceptance criteria             │
│  - APPLY: Execute with verification loops                   │
│  - UNIFY: Close loop, reconcile outcomes                    │
└──────────────────────────┬──────────────────────────────────┘

┌─────────────────────────────────────────────────────────────┐
│  5. REFLECT                                                 │
│  - Freedom to question                                      │
│  - Identify potential issues                                │
│  - Suggest improvements                                     │
│  - Update state                                             │
└─────────────────────────────────────────────────────────────┘

Component Architecture

┌─────────────────────────────────────────────────────────────┐
│                      Intent Router                          │
│  - Classifies requests                                      │
│  - Routes to strategy                                       │
│  - Manages confidence thresholds                            │
└──────────────────────────┬──────────────────────────────────┘

┌─────────────────────────────────────────────────────────────┐
│                    Context Manager                          │
│  - Loads project-specific rules                             │
│  - Manages context compression                              │
│  - Optimizes for position bias                              │
└──────────────────────────┬──────────────────────────────────┘

┌─────────────────────────────────────────────────────────────┐
│                     Skill System                            │
│  - 20 specialized skills                                    │
│  - Dynamic skill loading                                    │
│  - Skill chain orchestration                                │
└──────────────────────────┬──────────────────────────────────┘

┌─────────────────────────────────────────────────────────────┐
│                   PAUL Framework                            │
│  - Plan-Apply-Unify loops                                   │
│  - Mandatory loop closure                                   │
│  - Acceptance criteria                                      │
└──────────────────────────┬──────────────────────────────────┘

┌─────────────────────────────────────────────────────────────┐
│                  CARL Quality Gates                         │
│  - Context augmentation                                     │
│  - Quality gate enforcement                                 │
│  - Reinforcement layer                                      │
└─────────────────────────────────────────────────────────────┘

Key Concepts

Intent Routing

The first decision point in every request:

ComplexityStrategyLatency
LowDirect response1-2s
MediumSingle skill5-15s
HighSkill chain15-45s
Very HighRLM orchestration45-120s

Learn more →

Context Management

Position-aware loading of project rules:

  • Critical info at start/end (100% weight)
  • Supporting details (75% weight)
  • Less important in middle (50% weight)

Learn more →

PAUL Methodology

Plan-Apply-Unify Loop — Structured development:

  1. PLAN — Define objectives and acceptance criteria (Given/When/Then)
  2. APPLY — Execute tasks sequentially with verification
  3. UNIFY — Reconcile plan vs actual, update state, log decisions

Never skip UNIFY!

Quality over speed-for-speed's-sake. In-session context over subagent sprawl.

Learn more →

CARL Quality Gates

Context Augmentation & Reinforcement Layer — Dynamic rule loading:

  • Context augmentation — Loads rules just-in-time based on domain
  • Quality gate checks — Enforces with priority (Critical > High > Medium > Low)
  • Reinforcement layer — Blocks critical violations, warns on high

Three Domains:

  1. PAUL Domain — Loop enforcement, boundary protection
  2. Development Domain — Code quality, error handling, testing
  3. Projects Domain — Documentation, version handling

Learn more →

Model-Aware Editing

Dynamic edit format selection:

ModelFormat
Claude, Mistralstr_replace
Geministr_replace_fuzzy
GPTapply_patch
Grok, GLMhashline

Learn more →

Installation Paths

Tachikoma supports two installation methods:

TypePathPrecedence
Localcwd/.opencodeHigher
Global~/.config/opencodeLower

Discovery: Tachikoma checks both locations, with local taking precedence.

Design Principles

1. Token Efficiency

Optimize for tokens-per-task, not tokens-per-request:

  • Reference context modules instead of duplicating
  • Compress context at 70-80% utilization
  • Use structured summaries
  • Write large outputs to files

2. Quality First

Prioritize correctness over speed:

  • Verification loops for critical tasks
  • PAUL methodology for structure
  • CARL quality gates for enforcement
  • Model-aware editing for precision

3. Research-Backed

Base decisions on peer-reviewed research:

  • Position bias awareness
  • Cost-aware routing
  • Generator-Verifier-Reviser patterns
  • Adaptive chunking for large contexts

4. User Agency

Maintain human oversight:

  • Clarification when confidence < 0.7
  • Freedom to question in reflection phase
  • Explicit approval for destructive operations
  • Clear documentation of changes

Quick Reference

NeedLink
Install TachikomaGetting Started
Understand routingIntent Routing
Use skillsSkill Execution
Apply PAULPAUL Methodology
Enforce qualityCARL Quality Gates
Understand researchResearch Overview
View internalsInternals

Next Steps

Released under the MIT License.