Agent Skillskenn-io/roborev › roborev-review-branch

roborev-review-branch

GitHub

用于在指定分支上请求代码审查并展示结果。支持指定基准分支、审查类型及安全面板配置,自动验证 Git 引用有效性,处理沙箱权限问题,并在审查完成后向用户呈现最终报告。

internal/skills/codex/roborev-review-branch/SKILL.md kenn-io/roborev

触发场景

用户显式调用 $roborev-review-branch 需要检查当前分支所有提交的代码质量或安全漏洞

安装

npx skills add kenn-io/roborev --skill roborev-review-branch -g -y
更多选项

非标准路径

npx skills add https://github.com/kenn-io/roborev/tree/main/internal/skills/codex/roborev-review-branch -g -y

不安装直接使用

npx skills use kenn-io/roborev@roborev-review-branch

指定 Agent (Claude Code)

npx skills add kenn-io/roborev --skill roborev-review-branch -a claude-code -g -y

安装 repo 全部 skill

npx skills add kenn-io/roborev --all -g -y

预览 repo 内 skill

npx skills add kenn-io/roborev --list

SKILL.md

Frontmatter
{
    "name": "roborev-review-branch",
    "description": "Use only when the user explicitly invokes $roborev-review-branch"
}

roborev-review-branch

Request a code review for all commits on the current branch and present the results.

Usage

$roborev-review-branch [--base <branch>] [--type security|design] [--panel <name>|none]

Explicit invocation only

Invocation must be explicit: literal personal $roborev-review-branch, plugin $roborev:roborev-review-branch, or structured Codex skill selection. Requests such as “review this branch” without one of these explicit mechanisms must use native behavior and must not run roborev.

Sandbox access

roborev uses a local daemon. If a command fails with permission denied, the sandbox may be blocking access to its loopback port or Unix socket. Retry the same command with sandbox_permissions: "require_escalated". Do not start or restart the daemon because a sandboxed status probe cannot reach it.

When NOT to invoke this skill

Do NOT invoke this skill when the user is presenting or pasting existing review results. Messages that contain review findings, verdicts, or summaries are outputs — not requests to start a new review.

IMPORTANT

This skill requires you to execute bash commands to validate inputs and run the review. The task is not complete until the review finishes and you present the results to the user.

These instructions are guidelines, not a rigid script. Use the conversation context. Skip steps that are already satisfied. Defer to project-level AGENTS.md instructions when they conflict with these steps.

Instructions

When the user invokes $roborev-review-branch [--base <branch>] [--type security|design] [--panel <name>|none]:

1. Validate inputs

If a base branch is provided, use the base-branch command snippet below; it stores and validates the ref before invoking roborev review.

The snippet recovers one case on its own: when the ref names a configured remote followed by a branch (for example upstream/main) and that remote-tracking ref has not been fetched into this worktree yet, it fetches that one branch from that one remote and re-validates. It searches configured remote names from the longest slash-delimited prefix, so remote names that contain slashes are supported. Every other unresolvable ref is still rejected, and no fetch is attempted for a ref with no configured remote prefix.

If validation fails, inform the user the ref is invalid and report the git error. Do not proceed.

2. Build and run the command

Construct and execute the review command:

If no base branch is specified, run:

roborev review --branch --wait [--type <type>] [--panel <name>|none]

If a base branch is specified, run:

read -r branch <<'ROBOREV_REF'
<branch>
ROBOREV_REF
if ! git rev-parse --verify --quiet --end-of-options "$branch" >/dev/null; then
  remote=
  remote_branch="${branch##*/}"
  remote_candidate="${branch%/*}"
  while :; do
    if [ "$remote_candidate" != "$branch" ] && git config --get "remote.$remote_candidate.url" >/dev/null; then
      remote="$remote_candidate"
      break
    fi
    case "$remote_candidate" in
      */*)
        remote_branch="${remote_candidate##*/}/$remote_branch"
        remote_candidate="${remote_candidate%/*}"
        ;;
      *)
        break
        ;;
    esac
  done
  if [ -n "$remote" ]; then
    git check-ref-format --branch "$remote_branch" >/dev/null || exit 1
    git fetch --quiet --refmap= -- "$remote" "refs/heads/$remote_branch:refs/remotes/$remote/$remote_branch" || exit 1
    branch="refs/remotes/$remote/$remote_branch"
  fi
  git rev-parse --verify --end-of-options "$branch" >/dev/null || exit 1
fi
roborev review --branch --wait --base "$branch" [--type <type>] [--panel <name>|none]
  • If --base is specified, include it (otherwise auto-detects the base branch)
  • If --type is specified, include it
  • If --panel <name> is specified, include it (fans out to the named config panel); --panel none forces a single-agent review

The --wait flag blocks until the review completes.

3. Present the results

If the command output contains an error (e.g., daemon not running, repo not initialized, review errored), report it to the user. Suggest roborev status to check the daemon, roborev init if the repo is not initialized, or re-running the review.

Otherwise, present the review to the user:

  • Show the verdict prominently (Pass or Fail)
  • If there are findings, list them grouped by severity with file paths and line numbers so the user can navigate directly
  • If the review passed, a brief confirmation is sufficient

Panels (multi-reviewer reviews)

If you pass --panel <name>, or a default_panel is configured for explicit reviews, the review fans out to a panel of reviewers. In that case the Enqueued job <id> is the synthesis (parent) job that aggregates them, and its verdict and findings are the synthesized result across the whole panel. Present that synthesized verdict/findings, and offer fix on that parent id — never an individual reviewer. roborev show prints a one-line reviewers summary (e.g. 3 reviewers: bug P, security F) for a synthesis job. --panel none forces a single-agent review, and automatic post-commit hook reviews stay single-agent regardless of default_panel.

4. Offer next steps

If the review has findings (verdict is Fail), offer to address them:

  • "Would you like me to fix these findings? You can run $roborev-fix <job_id>"

Extract the job ID from the review output to include in the suggestion. Look for it in the Enqueued job <id> for ... line or in the review header. For a panel review this id is the synthesis parent.

If the review passed, confirm the result and do not offer $roborev-fix.

Examples

Default branch review:

User: $roborev-review-branch

Agent:

  1. Executes roborev review --branch --wait
  2. Presents the verdict and findings grouped by severity
  3. If findings exist: "Would you like me to address these findings? Run $roborev-fix 1042"
  4. If passed: "Branch review passed with no findings."

Security review against a specific base:

User: $roborev-review-branch --base develop --type security

Agent:

  1. Validates: git rev-parse --verify --end-of-options "develop"
  2. Executes roborev review --branch --wait --base develop --type security
  3. Presents the verdict and findings
  4. If findings exist: "Would you like me to address these findings? Run $roborev-fix 1043"

See also

  • $roborev-design-review-branch — shorthand for $roborev-review-branch --type design
  • $roborev-fix — fix a review's findings in code
  • $roborev-review — review a single commit

版本历史

  • 5f9fb4a 当前 2026-09-09 02:48

    修复了审查分支验证中未拉取的远程引用问题,现在会在验证失败前尝试从配置的远程仓库拉取缺失的分支引用。

  • 8bd3dcb 2026-09-03 04:19

    将技能文件中的引用从 CLAUDE.md 更新为 AGENTS.md,以适配 Codex 的项目策略文件加载机制。

  • 3148b42 2026-08-16 02:32

    新增沙箱环境下的守护进程发现机制,解决因权限拒绝导致的误判,增加 Unix socket 回退支持。

  • dec53d0 2026-07-24 12:15

同 Skill 集合

internal/skills/claude/roborev-design-review-branch/SKILL.md
internal/skills/claude/roborev-design-review/SKILL.md
internal/skills/claude/roborev-fix/SKILL.md
internal/skills/claude/roborev-lookahead-review-branch/SKILL.md
internal/skills/claude/roborev-lookahead-review/SKILL.md
internal/skills/claude/roborev-refine/SKILL.md
internal/skills/claude/roborev-respond/SKILL.md
internal/skills/claude/roborev-review-branch/SKILL.md
internal/skills/claude/roborev-review/SKILL.md
internal/skills/claude/roborev-snooze/SKILL.md
internal/skills/codex/roborev-design-review-branch/SKILL.md
internal/skills/codex/roborev-design-review/SKILL.md
internal/skills/codex/roborev-fix/SKILL.md
internal/skills/codex/roborev-lookahead-review-branch/SKILL.md
internal/skills/codex/roborev-lookahead-review/SKILL.md
internal/skills/codex/roborev-refine/SKILL.md
internal/skills/codex/roborev-respond/SKILL.md
internal/skills/codex/roborev-review/SKILL.md
internal/skills/codex/roborev-snooze/SKILL.md
internal/skills/droid/roborev-design-review-branch/SKILL.md
internal/skills/droid/roborev-design-review/SKILL.md
internal/skills/droid/roborev-fix/SKILL.md
internal/skills/droid/roborev-lookahead-review-branch/SKILL.md
internal/skills/droid/roborev-lookahead-review/SKILL.md
internal/skills/droid/roborev-refine/SKILL.md
internal/skills/droid/roborev-respond/SKILL.md
internal/skills/droid/roborev-review-branch/SKILL.md
internal/skills/droid/roborev-review/SKILL.md
internal/skills/droid/roborev-snooze/SKILL.md
internal/skills/grok/roborev-design-review-branch/SKILL.md
internal/skills/grok/roborev-design-review/SKILL.md
internal/skills/grok/roborev-fix/SKILL.md
internal/skills/grok/roborev-lookahead-review-branch/SKILL.md
internal/skills/grok/roborev-lookahead-review/SKILL.md
internal/skills/grok/roborev-refine/SKILL.md
internal/skills/grok/roborev-respond/SKILL.md
internal/skills/grok/roborev-review-branch/SKILL.md
internal/skills/grok/roborev-review/SKILL.md
internal/skills/grok/roborev-snooze/SKILL.md

元信息

文件数
0
版本
5f9fb4a
Hash
9b0f60c6
收录时间
2026-07-24 12:15

首页 - Wiki
Copyright © 2011-2026 iteam. Current version is 2.155.2. UTC+08:00, 2026-09-10 11:41
浙ICP备14020137号-1 $访客地图$