Architecture¶

Devran AI Kit v5.2.8 is an engineered framework with a 43-module runtime engine, 26 agents, 39 skills, 40 commands, 25 workflows, and 15 governance rules.

Architecture Overview¶

┌─────────────────────────────────────────────────────────────────────┐
│                      DEVRAN AI KIT v5.2.8                       │
├─────────────────────────────────────────────────────────────────────┤
│                                                                      │
│  ┌─────────────────┐  ┌─────────────────┐  ┌─────────────────┐      │
│  │   26 AGENTS     │  │   40 COMMANDS   │  │   39 SKILLS     │      │
│  │                 │  │                 │  │                 │      │
│  │ • Architect     │  │ • /plan         │  │ • api-patterns  │      │
│  │ • Mobile Dev    │  │ • /implement    │  │ • architecture  │      │
│  │ • DevOps        │  │ • /verify       │  │ • clean-code    │      │
│  │ • DB Architect  │  │ • /deploy       │  │ • testing       │      │
│  │ • Security      │  │ • /debug        │  │ • docker        │      │
│  │ • + 21 more     │  │ • + 35 more     │  │ • + 34 more     │      │
│  └─────────────────┘  └─────────────────┘  └─────────────────┘      │
│                              │                                       │
│           ┌──────────────────┴──────────────────┐                   │
│           ▼                                      ▼                   │
│  ┌─────────────────────────────────────────────────────────┐        │
│  │                    25 WORKFLOWS                          │        │
│  │  /brainstorm • /create • /debug • /deploy • /enhance    │        │
│  │  /implement • /orchestrate • /plan • /pr • /preflight    │        │
│  │  /preview • /quality-gate • /retrospective • /review     │        │
│  │  /project-status • /test • /ui-ux-pro-max • /upgrade     │        │
│  └─────────────────────────────────────────────────────────┘        │
│                              │                                       │
│           ┌──────────────────┴──────────────────┐                   │
│           ▼                                      ▼                   │
│  ┌─────────────────┐                   ┌─────────────────┐          │
│  │     RULES       │                   │     HOOKS       │          │
│  │  (Governance)   │                   │  (8 Lifecycle)  │          │
│  └─────────────────┘                   └─────────────────┘          │
│                                                                      │
│  ┌─────────────────────────────────────────────────────────┐        │
│  │              RUNTIME ENGINE (43 modules)                 │        │
│  │  workflow-engine • session-manager • task-governance      │        │
│  │  agent-reputation • self-healing • marketplace            │        │
│  │  plugin-system • identity • command-bridge • worktree     │        │
│  └─────────────────────────────────────────────────────────┘        │
│                                                                      │
└─────────────────────────────────────────────────────────────────────┘

Component Breakdown¶

Agents (26)¶

Specialized sub-agents that handle delegated tasks with focused expertise.

Core: Architect, Planner, Code Reviewer, TDD Specialist
Domain: Mobile, Frontend, Backend, Database, DevOps, Security, Performance
Support: Documentation, Build Error Resolver, Refactorer, Explorer, Knowledge
Language: TypeScript Reviewer, Python Reviewer, Go Reviewer
Autonomy: Sprint Orchestrator, Reliability Engineer, PR Reviewer

Commands (40)¶

Slash commands for quick execution of common operations.

Planning: /plan, /implement, /project_status, /setup
Development: /build, /fix, /debug, /refactor, /cook
Quality: /verify, /code-review, /security-scan, /perf

Skills (39)¶

Domain expertise modules that extend AI capabilities.

Operational: verification-loop, continuous-learning, strategic-compact, eval-harness, context-budget, production-readiness, project-docs-discovery
Orchestration: intelligent-routing, parallel-agents, behavioral-modes, mcp-integration
Planning: plan-writing (tiered quality schema), plan-validation (completeness scoring), brainstorming, research-methodology (multi-source evidence protocol, competitive analysis)
Domain: api-patterns, architecture, clean-code, database-design, testing-patterns, pr-toolkit, security-practices, and 14 more

Workflows (25)¶

Complete development lifecycles for multi-step processes.

/brainstorm, /create, /debug, /deploy, /enhance, /implement, /orchestrate
/plan, /pr, /pr-review, /pr-fix, /pr-merge, /pr-split, /preflight, /preview, /quality-gate, /retrospective, /review
/project-status, /help-kit, /test, /ui-ux-pro-max, /upgrade

All 25 workflows include Scope Filter tables, Argument Parsing, Failure Output templates, and measurable Completion Criteria.

Runtime Engine (43 Modules)¶

Node.js runtime modules that enforce governance, manage state, and provide platform features.

Phase	Modules
Phase 1 — Foundation	`workflow-engine`, `session-manager`, `verify`, `updater`, `error-budget`, `io`, `constants`, `logger`
Phase 2 — Runtime	`workflow-persistence`, `workflow-events`, `agent-registry`, `loading-engine`, `hook-system`, `task-model`, `learning-engine`
Phase 3 — Collaboration	`identity`, `task-governance`, `skill-sandbox`, `conflict-detector`, `security-scanner`, `plugin-system`, `plugin-verifier`
Phase 4 — Platform	`agent-reputation`, `engineering-manager`, `self-healing`, `marketplace`, `cli-commands`, `commands/init`
Phase 5 — IDE & Bridge	`command-bridge`, `ide-generator`, `project-ide-generator`, `worktree`, `config-validator`, `doc-discovery`, `doc-generator`
Phase 6 — Integrations	`telegram-sync`, `telegram-menu-guard`, `market-research`, `circuit-breaker`, `rate-limiter`, `quality-score`, `decision-validator`, `onboarding-engine`

Rules & Hooks¶

Rules (15): Governance constraints — security, coding-style, testing, git-workflow, documentation, architecture, performance, accessibility, data-privacy, sprint-tracking, workflow-standards, quality-gate, agent-upgrade-policy, market-awareness, doc-freshness
Hooks (9): Event-driven automation (session-start, session-end, pre-commit, phase-transition, sprint-checkpoint, secret-detection, plan-complete, task-complete, onboarding-complete)

Directory Structure¶

.agent/
├── agents/               # 26 specialized agents
├── commands/             # 40 slash commands
├── skills/               # 39 capability modules
├── workflows/            # 25 process templates
├── engine/               # Autonomy Engine configs
├── hooks/                # Event automation
├── rules/                # Governance rules
├── checklists/           # Verification checklists
├── templates/            # Feature templates
└── decisions/            # ADR system

lib/                      # Runtime Engine (43 modules)
├── workflow-engine.js    # State machine enforcement
├── task-governance.js    # Locking, audit trail, decision timeline
├── agent-reputation.js   # Score tracking & rankings
├── self-healing.js       # CI failure detection & patch generation
├── marketplace.js        # Community skill marketplace
├── command-bridge.js     # Cross-IDE slash command bridge generation
├── worktree.js           # Git worktree support (.worktreeinclude, post-checkout)
├── io.js                 # Gitignore management (narrow, cleanup, addToGitignore, untrackKitArtifacts, isSharedMode)
├── telegram-sync.js      # Telegram bot menu synchronization
└── + 34 more modules     # Identity, plugins, hooks, registry...

tests/                    # 1037 tests (54 files)
├── unit/                 # Module tests
├── structural/           # Inventory + schema validation
├── integration/          # Cross-module tests
└── security/             # Injection scan + leakage detection