hedypamungkas
hedypamungkas/koboi-agent
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。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/hedypamungkas/koboi-agent --skill code-review命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
This skill should be used when the user asks to "check the GitHub issues", "triage the open issues", "assess and rank the issues", "what issues should we fix", "fix the top issues", or otherwise wants to go from the issue tracker to fixed, merged PRs. It runs the whole loop end-to-end without further prompting — verify each issue empirically against current code, rank by severity × effort, surface the top high/critical ones (TDD is the default method, not an option), and once the user confirms scope, automatically dispatch the fixes in parallel as isolated worktree workers, each shipping its own PR, plus a reproducible harness, CI verification, and a memory note.
This skill should be used when the user asks to "release koboi-agent", "bump version and publish", "create a new release", "push to PyPI", "tag and release", "ship version vX.Y.Z", or mentions PyPI publishing, GitHub releases, or GHCR container image publishing for the koboi-agent project. Encodes the full release sequence (merge → pre-check → bump → tag → PyPI + GHCR auto-publish → verify → GitHub release) plus the gotchas learned from releases v0.3.0–v0.4.3.
This skill should be used when the user asks to "align docs", "sync documentation", "update README", "update CLAUDE.md", "fix stale docs", "audit documentation", "ensure docs match code", or mentions any discrepancy between documentation files (README, CLAUDE.md, architecture.md, configs/CLAUDE.md, folder CLAUDE.mds, etc.) and the actual codebase. Spawns parallel exploration agents to map the code + audit every doc, then revises all stale documentation to match — compact, jargon-free, human-and-LLM-friendly. Gates new-doc creation via AskUserQuestion.
基于 SOC 职业分类
正在显示 SKILL.md
| name | code-review |
| description | Systematic code review focusing on security, quality, and performance |
| license | MIT |
| disable-model-invocation | true |
| disallowed-tools | shell |
When this skill is activated, review code systematically:
read_file to examine the code