oc-pr
评估和处理 OpenCreator GitHub PR,先审查实际功能与风险,对不能合入的 PR 用英文反馈,并将满足门禁的 PR 合并到 develop。用于查看、评审或处理 https://github.com/krillinai/OpenCreator/pulls;不用于合并 master 或发布。
Install / Use
npx skills add krillinai/OpenCreator --skill oc-prInstalls into whichever agent you are using.
SKILL.md
Installable skill definition
Quality Score
Category
Development & EngineeringSupported Platforms
Our assessment of oc-pr
oc-pr scores 89/100 on our quality scale, 487th of 1,999 Development & Engineering skills we index (top 25%).
Its SKILL.md is 3.3 KB long, well organised into 9 sections with 1 code example: a solid amount of guidance for an agent.
With 12,288 GitHub stars, it is one of the more widely adopted skills in the catalogue.
Maintenance, license and trust
- The repository was last updated yesterday, so oc-pr is actively maintained.
- It is released under the Apache-2.0 license, a permissive license that allows use, modification and commercial use with attribution.
- Its trust signals score 100/100, with no cautions. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.
Safety scan
No issues foundOur scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands.
Automated pattern scan on 2026-09-26. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.
oc-pr compared with similar skills
All 4 of these similar skills score higher than oc-pr; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| oc-pr (this skill)by krillinai | 89 | 12.3k | 1d ago | SKILL.md |
| Agent-Reachby Panniantong | 100 | 85.5k | 10d ago | CLAUDE.md |
| headroomby headroomlabs-ai | 100 | 73.8k | today | CLAUDE.md |
| ai-job-searchby MadsLorentzen | 100 | 44.0k | 5d ago | CLAUDE.md |
| claude-howtoby luongnv89 | 100 | 41.7k | today | CLAUDE.md |
Frequently asked questions
- How do I install oc-pr?
- Run
npx skills add krillinai/OpenCreator --skill oc-pr. The install tabs above show the steps for each supported agent. - Which AI agents does oc-pr work with?
- It is written for OpenAI Codex, as a SKILL.md file. Other agents that read the same format can often use it too.
- Is oc-pr safe to use?
- Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. It is Apache-2.0-licensed and scores 100/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is oc-pr still maintained?
- The repository was last updated yesterday, so oc-pr is actively maintained.
Skill content
View source on GitHubname: oc-pr description: 评估和处理 OpenCreator GitHub PR,先审查实际功能与风险,对不能合入的 PR 用英文反馈,并将满足门禁的 PR 合并到 develop。用于查看、评审或处理 https://github.com/krillinai/OpenCreator/pulls;不用于合并 master 或发布。
OpenCreator PR
处理 https://github.com/krillinai/OpenCreator/pulls 的实时开放 PR。遵守根目录 AGENTS.md,所有产品 PR 的目标分支必须是 develop。
先确定授权模式
- 用户说“查看、评估、review”时只读:不得提交 GitHub review、修改 PR、合并或推送。
- 用户说“处理 PR、反馈并合并、合并可用 PR”时可提交英文 review,并把满足门禁的 PR 合并到
develop。 - 不把处理 PR 的授权扩展为合并
master、强制管理员合并、创建 tag 或发布。
获取事实
- 使用 GitHub API/
gh pr list获取当前开放 PR;不得用本地陈旧refs/pr/*代替页面状态。 - 对每个 PR 读取 base/head、提交、文件、描述、checks、reviews、mergeability 和已有讨论。
- 网络慢时可设置
HTTP_PROXY、HTTPS_PROXY为http://127.0.0.1:7897;先确认gh auth status。 - base 不是
develop时不得原样合并。用英文要求 retarget,或在用户明确授权后调整 base。
评审顺序
先解释 PR 为用户增加或改变什么功能,再检查实现。不要只根据 CI 颜色或文件数量判断。
执行评审前读取 references/review-checklist.md。重点确认业务正确性、失败/取消/重试、数据与历史版本、凭据隐私、协议边界、Web/Desktop 一致性和 Creator Panel 架构。
测试证据按风险选择:先运行最接近的定向测试和受影响包 typecheck;只有共享层或真实交互无法局部覆盖时扩大。不要为评审普通 PR 自动运行 Desktop 打包。
决策
APPROVE:功能正确、边界合理、必要验证通过、无阻断问题。COMMENT:有非阻断建议,不影响合入。REQUEST_CHANGES:存在可复现功能错误、安全/数据风险、架构违规或必要验证缺口。BLOCKED_BY_CONFLICTS:与最新develop冲突,必须更新分支并重新验证。
不能合入时,按 reference 中的格式用英文提交具体 review:说明当前行为、用户影响、证据和期望修复。不要只写 “tests missing” 或泛泛的风险。
合并到 develop
处理模式下,先验证合并结果,再修改远端 develop。只有以下条件全部满足才合并:
- base 是
develop,PR 非 draft。 - GitHub 报告 mergeable,且没有未解决的阻断 review。
- required checks 通过;无 checks 的高风险改动不得直接合并。
- 功能评审没有阻断问题,必要的定向验证通过。
- 合并前重新获取
origin/develop,避免基于过期状态判断。 - PR 与最新
origin/develop形成的独立 merge commit 已通过$oc-verifyL1;多个重叠 PR 或共享层改动使用 L2。
本地合并和验证
- 要求当前 Git 已跟踪文件没有未提交修改;不得把用户已有修改带入 PR merge commit。获取最新
origin/develop,记录 base SHA 和 PR head SHA。 - 从精确的
origin/develop创建仅供本次 PR 使用的本地集成分支,例如review/pr-<number>-merge。不创建 worktree,也不并行处理其他 PR。 - 使用
git merge --no-ff <pr-head>创建一个只包含该 PR 的独立 merge commit,提交信息标明 PR 编号和来源分支。发生冲突时执行git merge --abort,切回develop,将 PR 标记为BLOCKED_BY_CONFLICTS,并用英文要求提交者更新分支。除非用户明确授权代为解决,否则不得在维护者分支解决冲突。 - 创建 merge commit 后完成本轮
$oc-pr,不得自动调用$oc-verify。向用户报告 PR 功能、评审结论、base SHA、PR head SHA、merge commit,以及建议的验证级别和预计范围,然后暂停等待确认。
显式验证交接
交接提示必须明确说明下一步会做什么,例如:
PR #<number> 已完成评审并生成本地 merge commit <sha>,尚未更新 develop。
建议继续执行 $oc-verify L1(或 L2),预计验证:<测试范围>。
验证通过后将 fast-forward 并 push develop;失败则保持远端不变。是否继续?
只有用户明确确认后,才读取并执行 $oc-verify。确认同时授权完成提示中已明确写出的后续动作:验证通过后继续 fast-forward 和 push develop;如果提示没有说明 push,则验证后必须再次询问,不得自行推送。
验证失败或仅得到 PARTIALLY_VERIFIED 时切回 develop,不得 fast-forward、push 或改写远端;保留失败集成分支供诊断,是否删除由后续明确操作决定。只有 VERIFIED 或适用于该精确 merge commit 的 VERIFIED_BY_CI 才能继续合并;L3 的 MASTER_SOURCE_READY 不属于普通 PR 必需条件。
验证通过后再次 fetch,并确认 origin/develop 仍是记录的 base SHA。若已前进,当前 merge commit 失效:切回 develop,从新的 origin/develop 重新创建独立集成分支、merge commit 并验证,不得直接 rebase 后复用旧结果。
base 未变化时,切回本地 develop,确保它可以 --ff-only 到已验证 merge commit,然后 push develop。不得调用会重新生成另一个 merge/squash commit 的 GitHub 合并方式;远端必须收到本机已经验证的精确 commit。push 后确认 origin/develop、本地 develop 和已验证 commit 完全一致,并确认 GitHub PR 已关闭或标记 merged。
多个 PR 必须逐个处理:每个 PR 一个 --no-ff merge commit,每次验证通过并 push 后刷新 origin/develop,再处理下一个。不要在一个提交里混合多个 PR;后续 PR 出现回归时通过 revert 对应 merge commit 恢复,不对共享 develop reset 或 force push。
交付
逐个列出功能、关键发现、GitHub 决策、merge commit 和合并后验证。未执行的真实 Provider、跨平台或 Packaged App 验证必须明确说明。
Related Skills
Agent-Reach
85.5kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
headroom
73.8kCompress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
ai-job-search
44.0kThe job search that runs on your machine. AI job application framework built on Claude Code: evaluate postings, tailor CVs, write cover letters, prep interviews. Fork it and own it.
claude-howto
41.7kA visual, example-driven guide to Claude Code — from basic concepts to advanced agents, with copy-paste templates that bring immediate value.
Languages
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.
