菜单
Use when kagura-engineer:doctor reports a blocked or failing environment — scaffolds missing config, installs and authenticates harness prerequisites, and generates Memory Cloud MCP config by shelling out to `kagura-engineer setup`, then re-checks with doctor.
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
Use before any kagura-engineer run/review, or when the harness reports a blocked environment — diagnoses the local setup (git, selected brain CLI, gh, ollama, memory backend, and gh-issue-driven plugin) by shelling out to `kagura-engineer doctor` and reports what must be fixed.
Use to measure whether memory grounding actually improves PR quality — shells out to `kagura-engineer eval` with one or more issue numbers, running the SAME fixed issue set in two arms (recall ON vs OFF) and printing an A/B table. HARNESS — high cost; it runs the full loop twice per issue and mutates the repo. Confirm with the user before launching.
Use to review a git ref, branch, or PR with the local kagura-engineer reviewer — shells out to `kagura-engineer review [target]`, returning a structured verdict and findings. With `--fix` it is a HARNESS that edits files and commits; without `--fix` it is read-only.
Use to drive a single GitHub issue to a pull request with the kagura-engineer harness — shells out to `kagura-engineer run` with an issue number (guard → recall → worktree → start → implement → ship → persist). HARNESS — this mutates the repo, creates a PR, and spends model budget; confirm with the user before launching.
Use to bootstrap a fresh checkout for the kagura-engineer harness — scaffolds a commented repo.yaml template and adds it to .gitignore by shelling out to `kagura-engineer init`. Run this before setup/run/review when no repo.yaml exists yet. Idempotent and never overwrites.
| name | setup |
| description | Use when kagura-engineer:doctor reports a blocked or failing environment — scaffolds missing config, installs and authenticates harness prerequisites, and generates Memory Cloud MCP config by shelling out to `kagura-engineer setup`, then re-checks with doctor. |
Thin wrapper around the kagura-engineer setup CLI verb. It repairs the environment
that kagura-engineer:doctor diagnoses. It reimplements nothing — it shells out to the
installed CLI and surfaces the result.
Announce: "Using the kagura-engineer:setup skill to prepare the harness environment."
Usage: kagura-engineer:setup — no arguments. Add --fix <step> to repair one step, --dry-run to preview.
Note: when
repo.yamlis missing,setupscaffolds it automatically and reports the blank Cloud fields asNEEDS_USER. Fill in the backend choice and any required workspace/context ids, then re-run. Setup may prompt for credentials unless run with--no-input.
Discover config. Pass --config <path> if repo.yaml is not in CWD.
(Optional) Preview. For a dry run that changes nothing:
kagura-engineer setup --dry-run --json
Run setup:
kagura-engineer setup --json
To repair a single step only, use --fix <step> where <step> is one of
git, claude-code, headless-permissions, gh, ollama,
ollama-models, memory-cloud, memory-mcp. Add --full to the full run
or the memory-mcp step to also install memory hooks and skills; the default
generates .mcp.json only.
The headless-permissions step asks before merging the documented baseline
into .claude/settings.json, preserves unrelated settings, and then asks the
human to confirm workspace trust for both the repo and its
.kagura-runs/<repo>/ parent. It never edits ~/.claude.json; with
--no-input it makes no change and reports NEEDS_USER.
Exit-code contract: 0 = all steps ok/skipped · 1 = any step failed
(takes priority over 2) · 2 = a step needs user input with no failure
(or a config / unknown --fix error).
Surface the result. Print the ran / skipped / needs_user / failed
groups from the JSON, leading with failed and needs_user.
Re-check and hand off. Re-run the diagnosis to confirm the repair:
kagura-engineer doctor --json
When overall is ok/warn, hand back to the caller so it can proceed to
kagura-engineer:run or kagura-engineer:review.