soia-team/soia-open-skills / skills/soia-meta-sync-skills
soia-meta-sync-skills
Verified按明确范围把共享技能目录软链到项目或宿主,写入前先 dry-run。触发:技能同步、同步预览、把技能链到项目
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 5.1 KB |
| agents/openai.yaml | config | 262 B |
| assets/config.example.yml | config | 214 B |
| references/soia-managed-skills.md | markdown | 2.8 KB |
| references/source-rules.md | markdown | 2.0 KB |
| references/targets-and-confirmation.md | markdown | 3.6 KB |
| scripts/install_selection.py | script | 2.2 KB |
| scripts/sync_soia_skills.py | script | 34.7 KB |
references/targets-and-confirmation.md
View on GitHub# Targets and Confirmation
## Target Directory Matrix
| Target id | Tool | Directory | Notes |
|---|---|---|---|
| `pi` | Pi (pi-coding-agent) | `~/.pi/agent/skills` | User-level skill directory. |
| `soia` | SOIA AI | `~/.soia/skills` | Optional named target; select it explicitly when the customer uses it. |
| `workbuddy` | WorkBuddy | `~/.workbuddy/skills` | Optional named target; select it explicitly when the customer uses it. |
| `claude` | Claude Code | `~/.claude/skills` | User-level skill directory. |
| `codex` | OpenAI Codex | `~/.codex/skills` | Codex user skill directory. |
| `agy` | Antigravity CLI | `~/.gemini/antigravity-cli/skills` | Consumer Google-account successor target; independent from Gemini CLI. |
| `gemini` | Gemini CLI | `~/.gemini/skills` | Retained for Standard/Enterprise, API Key, and Vertex AI lanes. |
| `kimi` | Kimi CLI | `~/.kimi/skills` | User-level skill directory. |
| `opencode` | OpenCode | `~/.config/opencode/skill` | Directory name is singular `skill`. |
| `qwen` | Qwen Code | `~/.qwen/skills` | Actual command name is `qwen`, not `qwencode`. |
| `cursor` | Cursor | `~/.cursor/skills` | User-level skill directory. |
| `qoder` | QoderCLI | `~/.qoder/skills` | Keep synced even when subscription is inactive. |
| `copilot` | GitHub Copilot | `~/.copilot/skills` | Linked only when selected. |
| `windsurf` | Windsurf | `~/.codeium/windsurf/skills` | Linked only when selected. |
| `trae` | Trae | `~/.trae/skills` | Linked only when selected. For the CN edition, also inspect `~/.trae-cn`. |
| `openclaw` | OpenClaw | `~/.openclaw/skills` | Linked only when selected. |
Command-name truth:
- `agy` is the Antigravity CLI command. It replaces only Gemini CLI consumer
Google login and must not be aliased to `gemini`.
- `qwen` is the command name, not `qwencode`.
- `opencode` is the command name and supports `run`, `serve`, and `web`.
## Target Selection
- Do not assume any product-specific target. Global scope requires explicit `--targets`; without it the script returns `selection_required` and writes nothing.
- A target selected by id is linked from the source directory supplied for this run; the tool never infers a repository checkout.
- When proposing choices, existing known target dirs (and `agy` when the `agy` command exists) are candidates to show the customer, not defaults to write.
- Allow explicit known target ids, comma-separated ids, and custom paths.
- Create selected target dirs during execution if they do not exist.
- Reject a target path that resolves to the same path as the source skills root.
- Target entries are symlinks. If a target already has a same-name copied directory, remove that directory and replace it with a symlink after confirmation.
## Confirmation Compatibility
AskUserQuestion-capable clients:
- Present target choices as a multi-select question.
- Recommend only targets explicitly named by the customer or present in their user-owned configuration.
- Add a final continue/cancel confirmation after showing the sync plan.
- Use the client-provided "Other" option for custom target paths when available.
Codex Default or plain-text clients:
- Ask the user to reply with target ids or paths in text.
- Show the final plan and command.
- Wait for explicit confirmation before running a writing command.
- Treat `confirm`, `yes`, `y`, `执行`, `继续`, `可以`, or `go` as confirmation only when it clearly refers to the displayed plan.
- A request to "install/sync now" is not by itself write approval: show the dry-run impact first, then write only after the customer approves that complete plan (source, targets, action, delete/replace impact).
Never execute writes from an implied or stale confirmation.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |