ワンクリックで
candid-ship
Ship your changes - review, build, test, create PR, and optionally auto-merge
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
メニュー
Ship your changes - review, build, test, create PR, and optionally auto-merge
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Use when you have an output — copy, a doc, an answer, a plan, a prompt, any text artifact — and want it made better against your goal. Runs a critique→refine loop with configurable stop conditions and auto/review-each/interactive modes. Default is a single pass using the current session context to decide what "better" means.
Fast ship for low-risk changes — runs only the steps you explicitly enable in fastShip config
Use when the code works and you want a focused pass on improving approach, clarity, and quality — distinct from candid-review's defect hunt. Each suggestion includes a concrete before/after, tradeoffs, and confidence level.
Run candid-review in a loop until all issues are resolved, with configurable auto, review-each, or interactive modes and support for ignored issues
Use when reviewing code changes before commit or PR — configurable harsh or constructive review against Technical.md standards, with categorized issues, actionable fixes, todo tracking, and optional auto-commit.
Take a candid-chrome-qa findings JSON, let the user pick which findings to fix, and dispatch fixes — single batched PR, one PR per finding via Conductor deep links, local-only, or issues-only. Optionally files Linear (or other tracker) issues per finding before shipping.
| name | candid-ship |
| description | Ship your changes - review, build, test, create PR, and optionally auto-merge |
Orchestrate the full shipping workflow: review code via candid-loop, run install/build/tests, create a GitHub PR, optionally update an issue tracker and auto-merge. Aborts on any failure before PR creation.
candid-ship runs all configured steps by default — the user opts out with --skip-* flags. (Compare candid-fast-ship, which is opt-in.)
Execute these steps in order. The mechanics for each step are defined in skills/candid-ship/WORKFLOW.md. This file specifies the skip semantics unique to candid-ship.
Run the pre-flight checks from WORKFLOW.md → "Pre-Flight Checks".
Read config per WORKFLOW.md → "Load Configuration", extracting these ship.* fields plus mergeTargetBranches:
buildCommand, testCommand, installCommand, targetBranch, autoMerge, additionalPrompt, postMergeCommand, issueTracker (full sub-tree).
When loading from project config, output: Using ship settings from project config. From user config: Using ship settings from user config.
| Flag | Description | Default |
|---|---|---|
--auto-merge | Enable auto-merge | from config |
--no-auto-merge | Disable auto-merge | from config |
--skip-review | Skip candid-loop step | false |
--skip-install | Skip install step | false |
--skip-build | Skip build command | false |
--skip-tests | Skip test command | false |
--dry-run | Show plan without executing | false |
If both --auto-merge and --no-auto-merge are provided, --no-auto-merge wins.
After loading, run the targetBranch resolution and branch-state validation from WORKFLOW.md → "Resolve targetBranch" / "Validate Branch State".
Execute WORKFLOW.md → "Classify Diff Risk". Enforce the candid-ship escalation defined there: HIGH risk with --skip-review or --skip-tests aborts before the plan is shown. Display Diff risk: [class] ([signals]) in the plan box, directly below the Branch: line.
Calculate totalSteps = 1 (PR creation always runs) + number of optional steps that will run. An optional step counts when:
review: --skip-review is not setinstall: --skip-install is not set AND installCommand is setbuild: --skip-build is not set AND buildCommand is settests: --skip-tests is not set AND testCommand is setissueTracker: issueTracker.enabled is trueautoMerge: autoMerge is truepostMergeCommand: autoMerge is true AND postMergeCommand is setRenumber the displayed step list to skip rows for any disabled optional steps.
Render the plan per WORKFLOW.md → "Display Plan": header Candid Ship Plan, statuses [or SKIP] / [or SKIP — not configured] per the enablement rules above.
If additionalPrompt is set, append: Review context: "[additionalPrompt]".
If --dry-run: Output Dry run complete. No changes made. and exit.
Otherwise: Use AskUserQuestion: "Proceed with this shipping plan?" with options "Yes, ship it" / "No, cancel". On "No, cancel": exit with Ship cancelled.
Run in order: review, install, build, tests. For each: skip if its --skip-* flag is set → Skipping [step] (--skip-[step]); skip install/build/tests if the corresponding command is not configured → Skipping [step] (not configured); otherwise execute the matching WORKFLOW.md section ("Run Review (candid-loop)", "Install Dependencies", "Run Build", "Run Tests").
Always runs (read-only checks; not counted in totalSteps). Execute, in order, from WORKFLOW.md:
Always runs. Execute WORKFLOW.md → "Create Pull Request".
Execute WORKFLOW.md → "Update Issue Tracker". The skip conditions there map directly to ship's behavior. Notes:
teamPrefixes (DIS, ENG, DISC) reflect one Linear workspace. Edit ship.issueTracker.teamPrefixes in .candid/config.json to match your workspace's team keys.prompt encodes four invariants (single-issue, single-field, idempotent, no fallback search). Custom prompts must preserve "single issue" and "no fallback search" or the pre-call check will refuse the call.Skip if autoMerge is false → Auto-merge: disabled (manual merge required). Otherwise execute WORKFLOW.md → "Auto-Merge".
Skip if postMergeCommand is not configured → Skipping post-merge command (not configured). Skip if autoMerge is false → Skipping post-merge command (auto-merge disabled). Skip if auto-merge failed in Step 10 → Skipping post-merge command (auto-merge failed). Otherwise execute WORKFLOW.md → "Post-Merge Command".
Execute WORKFLOW.md → "Display Summary" with header Candid Ship Complete.
Add to .candid/config.json:
{
"version": 1,
"ship": {
"buildCommand": "npm run build",
"testCommand": "npm test",
"installCommand": "pnpm install",
"targetBranch": "stable",
"autoMerge": false,
"additionalPrompt": "Focus on security and ensure all API endpoints have auth middleware",
"postMergeCommand": "curl -X POST https://deploy.example.com/trigger",
"issueTracker": {
"provider": "linear",
"enabled": true,
"teamPrefixes": ["DIS", "ENG", "DISC"],
"state": "In Review"
}
}
}
Types, defaults, and validation live in skills/candid-review/CONFIG.md (ship section); WORKFLOW.md → "Load Configuration" summarizes the defaults.
Note:
issueTrackeris an optional integration. When omitted, the issue-tracker step is skipped silently — the rest of the ship runs unchanged. To request support for another tracker (Asana, Jira, GitHub Issues, etc.), open an issue at https://github.com/ron-myers/candid/issues.
Minimal — just PR creation:
{ "ship": { "targetBranch": "main" } }
Full pipeline with install, auto-merge, and Linear:
{
"ship": {
"installCommand": "pnpm install",
"buildCommand": "pnpm build",
"testCommand": "pnpm test",
"targetBranch": "stable",
"autoMerge": true,
"additionalPrompt": "Ensure error handling covers all async operations",
"postMergeCommand": "curl -X POST https://deploy.example.com/trigger",
"issueTracker": {
"provider": "linear",
"enabled": true,
"teamPrefixes": ["DIS", "ENG"],
"state": "In Review"
}
}
}
# Ship with all defaults from config
/candid-ship
# Force auto-merge
/candid-ship --auto-merge
# Quick ship — skip review, just install/build/test/PR
/candid-ship --skip-review
# Skip install (deps already up to date)
/candid-ship --skip-install
# See what would happen without doing it
/candid-ship --dry-run
# PR only
/candid-ship --skip-review --skip-install --skip-build --skip-tests
Fail-fast: any pre-PR failure aborts immediately; post-PR steps (issue tracker, auto-merge, post-merge) warn but never abort — the PR already exists. Set installCommand (e.g. npm ci) so build/test failures reflect real code issues, not missing dependencies.