d1d7cf8ef2
* refactor: rename opencode-config to configure-opencode
- command: .opencode/commands/opencode-config.md -> configure-opencode.md
- skill dir: .opencode/skills/opencode-config/ -> configure-opencode/
- command body: skill({name: "opencode-config"}) -> skill({name: "configure-opencode"})
- project-map/README.md: update command+skill entries
* refactor: rewrite configure-opencode skill for .opencode/ auto-discovery
- flip canonical rule: config/opencode.json + bind-mount -> .opencode/opencode.json (project-local, auto-discovered, zero env var)
- repo is the template users clone (git clone && opencode)
- drop: bind-mount as primary (-> optional Docker), OPENCODE_CONFIG_DIR as required (-> power-user/Docker), Windows/ADR-009, internal env vars (ANTIDETECT/CONTEX7/CLOUDFLARE -> generic {env:MY_API_KEY})
- rename paths: config/ -> .opencode/, slaid098/opencode -> slaid098/opencode-config
- preserve: MCP/provider/permission formats, top-level keys, findLast semantic, check-permissions.py guard
- project-map: update entry descriptions for renamed command+skill
* docs(handoff): add pr-12 handoff + ADR-007
* docs(handoff): rename pr-12 handoff/ADR to pr-27 (actual PR number)
---------
Co-authored-by: opencode-agent <agent@slaid098.dev>
5 lines
No EOL
491 B
Markdown
5 lines
No EOL
491 B
Markdown
---
|
|
description: Edit opencode.json config (MCP, providers, permissions, agents)
|
|
agent: build
|
|
---
|
|
Load the `configure-opencode` skill via `skill({name: "configure-opencode"})` and apply its canonical rules to the user's request about opencode.json changes (MCP servers, providers, permissions, agents, plugins). Follow the skill's rules strictly: always write to `.opencode/opencode.json` in slaid098/opencode-config repo (project-local, auto-discovered), never project-local in other repos. |