| name | kade-hq |
| description | Use when the user wants the KADE System itself, including Layer 1 and Layer 2 context loading, KADE session structure, HUMAN.md and kade/ overlays, or project initialization and handoff workflows. This packet-owned wrapper is installable into ~/.agents/skills, ~/.codex/skills, and ~/.claude/skills. |
kade-hq
This is the packet-owned launcher wrapper for kade-hq.
g-kade is the bridge skill that coordinates kade-hq plus gstack.
Use kade-hq when the task is specifically about KADE context, profiles, overlays, or session flow.
Startup
- Read repo-local instructions first.
- Detect whether a richer external
kade-hq or kade-headquarters skill is installed.
A richer install is one that contains runtime or companion content beyond this wrapper, for example templates/, references/, scripts/, HUMAN.md, or multiple sibling support folders.
Check these locations in order:
~/.codex/skills/kade-hq
~/.claude/skills/kade-hq
~/.agents/skills/kade-hq
~/.codex/skills/kade-headquarters
~/.claude/skills/kade-headquarters
~/.agents/skills/kade-headquarters
- repo-local
.codex/skills/kade-hq
- repo-local
.claude/skills/kade-hq
- repo-local
.agents/skills/kade-hq
- repo-local
.codex/skills/kade-headquarters
- repo-local
.claude/skills/kade-headquarters
- repo-local
.agents/skills/kade-headquarters
If you find a richer install, read the upstream SKILL.md there and use it.
If you do not find one, continue with this wrapper.
Packet Integration
Keep the layering simple and recognizable:
- Treat
~/.kade/HUMAN.md as Layer 1: who Kade is, preferences, style, and personal operating context.
- Treat
kade/AGENTS.md and kade/KADE.md as Layer 2: the project-local Kade overlay and handoff surface.
- Treat the root packet files and
.llm-wiki/config.json as Layer 3: workspace contract, search, memory, and MCP wiring.
- Treat
g-kade as the unifier surface and gstack as the execution-workflow surface.
When knowledge needs to persist:
- put session/handoff context in
kade/KADE.md
- put reusable procedural shortcuts in
llm-wiki-skills
- put stable repo knowledge in
wiki/
This keeps the Kade homage intact: human layer first, project layer second, harness layer third.
Install Flow
When bootstrapping a repo workspace, prefer:
python installers/install_g_kade_workspace.py --workspace <repo-root>
powershell -NoProfile -ExecutionPolicy Bypass -File .\installers\install_g_kade_workspace.ps1 --workspace <repo-root>
bash ./installers/install_g_kade_workspace.sh --workspace <repo-root>
After the workspace files are present, wire Claude and Codex lifecycle hooks:
python installers/wire_repo_agent_hooks.py --workspace <repo-root> --agents claude,codex --self-test
If invoked from the Codex plugin package instead of the packet checkout, use:
python plugins/llm-wiki-organizer/scripts/install_repo_hooks.py --workspace <repo-root> --agents claude,codex --self-test
This flow should leave these KADE surfaces in place:
~/.kade/HUMAN.md
kade/AGENTS.md
kade/KADE.md
- repo-local
kade-hq, gstack, and g-kade skill surfaces
Constraints
- Abstract and simplify by default; do not overcomplicate the KADE layers with unnecessary mechanism names in user-facing explanations.
- Preserve the Kade homage: keep the human/profile layer legible and first-class.
- Do not treat
g-kade as a replacement for kade-hq; it is only the bridge skill.
- Preserve a real existing
~/.kade/HUMAN.md; only seed or replace the exact legacy stub.
- Report whether you used this wrapper or a richer external
kade-hq install.