# Agentic Engineering Protocol

This file is a minimal activation stub loaded into every Kimi Code CLI session in this
repository. The full methodology body (delegation, risk classification, quality gates, git
workflow, and the two rules files) is NOT embedded here - it is trigger-loaded via the
`dinostack` skill (DS-185, mirroring the DS-143 change that moved the equivalent always-loaded
body out of Claude Code's `~/.claude/CLAUDE.md`).

**Note:** This file is auto-generated by `.kimi/build.sh`. Do not edit it directly.

## Activation

Activation mode/profile is read from `~/.claude/agentic-engineering.json` - the same shared
config file every adapter uses. Run `bash .kimi/install.sh` to configure it.

## Load the skill on trigger

Before acting on any software development task in this repository (code edits, debugging,
testing, deployment, architecture decisions, git operations, agent orchestration, code review,
refactoring, dependency management, project setup), load the `dinostack` skill:

- `/skill:dinostack` (full methodology), or `/skill:<command-name>` for a specific command
  (e.g. `/skill:ds-wrap`, `/skill:ds-skeptic`, `/skill:ds-implement-ticket`)
- or read `.kimi/skills/dinostack/SKILL.md` directly (local copy) or
  `~/.kimi/skills/dinostack/SKILL.md` (global install) if skill loading is unavailable

The skill embeds the full methodology body - it loads on trigger, not automatically at session
start. When in doubt, invoke it.