菜单
Run Alibaba's self-hosted Open Code Review (`ocr`) CLI on Git changes to produce line-level review comments. Use as a self-hosted supplemental/manual reviewer alongside the CR -> BugBot -> Greptile chain. Not part of the merge gate.
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
Generate a PM handoff prompt for context-window turnover. Captures static config, live GitHub state, in-flight thread state, and memory summary into a self-contained prompt for a fresh PM thread. Bootstraps `.claude/pm-config.md` on first run. Triggers on "pm-handoff", "handoff", "context turnover", "new pm thread", "thread turnover".
Scan open issues for staleness and suggest closures. Detects issues solved by merged PRs, no-activity issues (30+ days), superseded issues, and potential duplicates of already-closed issues. Presents recommendations for user confirmation — never auto-closes. Triggers on "pm-clean", "stale issues", "clean backlog", "close stale".
Active PM orchestrator — manages issue pipeline, tracks coding threads, ranks the open backlog (OKR-aware) against a business goal, and suggests next work. Cold-starts from GitHub state or resumes from a /pm-handoff prompt. Triggers on "pm", "project manager", "orchestrate", "what should I work on", "rank issues".
Re-scan the current repo and update `.claude/pm-config.md` with fresh infrastructure and architecture detection while preserving user-edited sections (Role, OKRs, Team, Notes, Dependency Rules, Workflow Rules), then run a stale-cleanup pass to prune long-abandoned worktrees and branches. Use this after major milestones, new service integrations, or significant directory restructuring. Triggers on "pm update", "update pm config", "refresh pm config", "rescan repo".
Capture-only thread mode for drafting and opening well-structured GitHub issues. Puts the thread into issue-capture mode — the only job is creating, editing, and closing issues (no implementation, no worktrees, no /start-issue). Always reflects before writing (1–3 scope/clarifying questions), creates canonical 6-section bodies with functional-first tone, runs dedup search, suggests labels, supports batch + cross-references, and prints the issue URL as the closing line of every create/update. Supports /update <N> <statement>, natural-language edit-in-place, and retract. Invoke as `/issue-maker [rapid-fire] [--export-prompt]`.
Manage the OKRs section of `.claude/pm-config.md`. Show current objectives, set new ones, or get AI-suggested updates based on recent work. OKRs drive `/pm` ranking. Use this whenever you want to view, set, or update project OKRs, objectives, key results, or goals. Triggers on "pm okr", "set okrs", "show okrs", "suggest okrs", "project goals", "objectives".
| name | open-code-review |
| description | Run Alibaba's self-hosted Open Code Review (`ocr`) CLI on Git changes to produce line-level review comments. Use as a self-hosted supplemental/manual reviewer alongside the CR -> BugBot -> Greptile chain. Not part of the merge gate. |
| triggers | ["ocr run","open code review","ocr review","self-hosted review","review with ocr"] |
| model | sonnet |
| allowed-tools | ["Read","Glob","Grep","Bash","WebFetch"] |
Invoke open-code-review (ocr) — Alibaba's open-source, Apache-2.0, self-hosted AI code-review CLI — as a supplemental reviewer. ocr reads a Git diff, selects/bundles files deterministically, runs an LLM agent against a configurable endpoint, and emits structured line-level comments.
This skill adapts the upstream skills/open-code-review/SKILL.md (Apache-2.0, author: alibaba) to this repo's conventions. Source of truth for flags: ocr review -h.
Scope: OCR is a manual / advisory reviewer here. It is not in the merge gate (
cr-merge-gate.md) and does not satisfy review requirements. Promotion into the chain is a separate, deferred decision — see.claude/reference/ocr-eval.md. Cost: OCR runs against the same Anthropic key as everything else, so its tokens count toward the daily quota. Bigocr reviewruns on large PRs can be expensive — prefer--from/--toor--commitover whole-reposcan.
which ocr || echo "NOT INSTALLED" # install: npm install -g @alibaba-group/open-code-review
ocr version
ocr llm test # must return connectivity OK before a real review
If ocr llm test fails, OCR has no usable LLM credential. Never invent, hardcode, or commit an API key. OCR reuses the existing Anthropic credential — guide the user to provide it via env var (preferred) rather than writing it to the config file:
# Provider mode (config has provider: anthropic) + standard sk-ant-* key:
export ANTHROPIC_API_KEY=<existing-key> # satisfies the x-api-key header path
Persistent (non-secret) config lives in ~/.opencodereview/config.json. The vetted setup for this fleet is:
ocr config set provider anthropic
ocr config set providers.anthropic.model claude-opus-4-8 # current lineup (Opus 4.8)
ocr config set providers.anthropic.auth_header x-api-key # required for sk-ant-* keys
Config-mode gotcha (OCR 1.6.1): when provider is set, OCR uses the provider-scoped resolver (providers.<name>.* + ANTHROPIC_API_KEY) and ignores the legacy llm.* / OCR_LLM_* keys and ANTHROPIC_AUTH_TOKEN. The provider model must be set via providers.anthropic.model (or --model); ANTHROPIC_MODEL env is not read in this mode.
Summarize the intent of the change (issue link, what it enables) into one or two sentences. Pass it via --background to improve review quality.
Always use --audience agent (summary only, no progress UI):
| User says | Command |
|---|---|
| "review my changes" / working copy | ocr review --audience agent -b "<context>" |
| "review this PR" / feature branch | ocr review --audience agent -b "<context>" --from main --to <branch> |
| "review commit abc123" | ocr review --audience agent -b "<context>" --commit abc123 |
| "what would be reviewed?" (dry run) | ocr review --preview |
Useful flags: --format json (machine-readable), --concurrency <n> (lower if rate-limited), --timeout <min>, --model <id> (override, e.g. a Sonnet-tier model for cheaper bulk passes), --rule <file.json> (custom rules).
Markdown/config-repo caveat: OCR's deterministic pipeline excludes files with unsupported extensions (e.g. .md) as unsupported_ext. In doc/config-heavy repos like this one, ocr review --preview will show most files excluded — OCR adds little here without a custom rule/tools config. It is most valuable on code-heavy (TS/Python/Go) repos.
Group every OCR comment by priority and report to the user:
If the user asked to "review and fix", apply fixes. If they asked only to "review", ask before changing anything.
OCR contacts only the configured provider endpoint (Anthropic API by default). It does not send code to any third-party review service. Verify with ocr review --preview (no network) and by confirming the provider config (providers.<name> base URL) in ~/.opencodereview/config.json.