Skip to content

Latest commit

 

History

History
77 lines (50 loc) · 3.98 KB

dev-kit.md

File metadata and controls

77 lines (50 loc) · 3.98 KB

dev.kit: The Thin Empowerment Layer

Domain: Foundations / Core Concept
Status: Canonical

Summary

dev.kit is a high-fidelity engineering interface designed to resolve the Drift between human intent and repository reality. It operates as a Thin Empowerment Layer (Grounding Bridge) that transforms a static codebase into a dynamic "Skill Mesh" accessible to humans and AI agents alike.

Core Philosophy

dev.kit is built on the principles of Context-Driven Engineering (CDE). It does not replace your existing tools; it orchestrates them to maintain a deterministic, context-aware engineering environment.

  • Non-Proprietary: Uses standard Markdown, YAML, and Shell scripts.
  • Deterministic: Every action is bounded by a validated CLI interface.
  • Agent-Ready: Provides native "Grounding" for LLMs, transforming them into configuration engines.

The Three Pillars of dev.kit

1. Grounding (The Bridge)

dev.kit provides the necessary context to ensure that every engineering action is grounded in the repository's truth. It audits the environment (dev.kit doctor) and synchronizes AI context (dev.kit ai sync).

2. Normalization (The Filter)

Chaotic user requests are filtered through a Normalization Boundary. Ambiguous intent is distilled into a deterministic workflow.md plan before any execution occurs.

3. Execution (The Engine)

Logic is executed through modular, standalone scripts and CLI commands. dev.kit ensures these run in a consistent, environment-aware context via environment.yaml.

Architecture: The Empowerment Layer

dev.kit distinguishes between Deterministic Functions (the programmatic logic) and AI Reasoning Skills (the dynamic intent resolution).

1. Deterministic Functions (The Engine)

These are hardened, predictable routines found in lib/commands/ and docs/skills/*/assets/. They provide the execution engine for the repository.

  • Example (Git Sync): The workflow.yaml and git_sync.sh logic that groups files and executes commits.
  • Example (Visualizer): The Mermaid templates and mmdc export logic that renders SVGs.
  • Role: Execute specific, bounded actions with high fidelity.

2. AI Reasoning Skills (The Brain)

These are the dynamic capabilities defined in SKILL.md. They use LLM reasoning to bridge unstructured intent with repository functions.

  • Example (Git Sync): Analyzing a set of changed files to determine the logical domains (docs, cli, core) and generate a meaningful commit message.
  • Example (Visualizer): Reading a README or source file to extract the logical process flow and map it to a specific Mermaid template.
  • Role: Interpret intent and orchestrate the engine.

The Skill Mesh

dev.kit treats the entire repository as a Skill. It dynamically discovers the mesh by scanning:

  • Internal Commands: Metadata-rich shell scripts in lib/commands/.
  • AI Reasoning Skills: Authoritative SKILL.md files in docs/skills/.
  • Functional Assets: Programmatic templates and configs managed by the engine.
  • Virtual Capabilities: Global environment tools (gh, npm, docker).

Primary Interfaces

  • dev.kit status: The "Engineering Brief." High-signal overview of health and active tasks.
  • dev.kit ai: The "Grounding Layer." Orchestrates AI integration and skill synchronization.
  • dev.kit sync: The "Drift Resolver." Atomic, domain-specific repository synchronization.
  • dev.kit task: The "Lifecycle Manager." Tracks intent from normalization to resolution.

📚 Authoritative References

The mission of dev.kit is grounded in the practical need for high-fidelity engineering empowerment:

UDX DevSecOps Team