菜单
Start a paude container session for fire-and-forget agent work. Use when creating a paude session, delegating a task to paude, running work in an isolated container, setting up a paude worktree, or launching pi/claude/gemini in podman.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Review all pending changes against repo conventions before committing
Mark files as human-reviewed with specific validation types
Harvest and locally merge changes from a paude container session. Use when pulling agent work from paude, running paude harvest, merging session commits, importing submodule changes from a container, or fixing failed harvest/submodule fetch.
Write a task spec file for a paude container agent. Use when delegating work to paude, preparing a paude prompt-file, writing fire-and-forget agent tasks, or scoping submodule work for harvest.
Diagnose a running or stalled paude container session. Use when paude seems stuck, harvest returned empty, checking paude status, debugging 403 proxy blocks, or deciding wait vs harvest vs reset.
Create or update a project BRIEF.md — scaffolds planning structure for work that crosses the 5-file / multi-session threshold
| name | paude-launch |
| description | Start a paude container session for fire-and-forget agent work. Use when creating a paude session, delegating a task to paude, running work in an isolated container, setting up a paude worktree, or launching pi/claude/gemini in podman. |
| argument-hint | <session-name> [--slug <slug>] [--spec <absolute-path>] [--agent pi|claude|gemini] |
| allowed-tools | Read Write Shell Glob Grep |
<essential_principles>
main while other sessions may run.--prompt-file paths must be absolute — cwd changes inside worktrees.--git is required for harvest; without it the container workspace is empty.paude connect <name>) is separate from git branch slug — record both.</essential_principles>
- Decision gate: `references/decision-gate.md` - Flags: `references/create-flags.md` - Worktrees: `rules/git-worktrees.md` - Task specs: `.agents/skills/paude-spec/SKILL.md` - Narrative: `docs/ai-engineering/paude-getting-started.md`Read references/decision-gate.md. If paude is not the right tool, say so and stop.
From $ARGUMENTS:
<session-name> — required--slug <slug> — default: ask or derive from spec filename--spec <path> — task spec; if missing, invoke paude-spec first or ask--agent — default pi for this workspaceIf no spec file, run or point user to paude-spec before continuing.
Spec must include session name, harvest branch slug, and harvest-prep instructions.
From host workspace root (on current main):
git fetch origin && git pull origin main
git worktree add worktrees/<slug> -b <slug>
If worktree exists, ask: resume in existing worktree or new slug?
cd worktrees/<slug>
paude create <session-name> --git --yolo --agent <agent> \
--prompt-file <absolute-path-to-spec>
Read references/create-flags.md for domains, fork image, optional flags.
paude status <session-name>
Confirm: running, project path matches worktree.
Report:
paude status, paude waitpaude-harvest <session-name> --slug <slug> --merge<success_criteria>
--git sync from correct worktree