Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add XRenSiu/claude-code-forge --skill prgit clone --depth 1 https://github.com/XRenSiu/claude-code-forgeWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/xrensiu/claude-code-forge/pr)<a href="https://agentmods.dev/skills/xrensiu/claude-code-forge/pr"><img src="https://agentmods.dev/badge/skills/xrensiu/claude-code-forge/pr.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00274 | $0.02773 |
| Opus 5 | $0.00137 | $0.01386 |
| Sonnet 5 | $0.00055 | $0.00555 |
| Haiku 4.5 | $0.00027 | $0.00277 |
Grade A, and why
pr scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured today.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 114 lines — stays where its author put it; the contents beside it link to each section on GitHub.
pr — 一份能被机器检、能被人按图索骥的 PR
产物是一个 GitHub PR:body 按 assets/pr_template.md 的段落序呈现(产物序,出口可检)。本文件
写:PR 在流水线里是什么、什么算写对、原语与出口、建之前的门。先 rebase 还是先写 body、要不要拆
是你的份额——除非下面说了它是不可逆的。
缺口(Judgment + Capability + Control)
deletion 测试:撤掉本 skill,引擎 git push && gh pr create --fill:标题是最后一条 commit,body
是 commit 列表。reviewer 不知道范围在哪、AC 有没有被覆盖、怎么验证;/review-loop 判"越界"没有
基准;1200 行的 diff 一次扔过去。缺的是 body 的产物序判据、体量判据、以及把这些编译掉的预门。
世界(Σ)
- PR body 的范围声明是下游的基准:/review-loop 判"越界类"评论对照它;/pr-review 判 B 档 "公共 API 变更"对照它。写得越具体,后面判定越省力。
- AC → 证据映射是 A 档验收的输入:每条 mechanical AC 对应一个可复现的证据(测试名 + 命令 +
结果、或 evaluation_result 路径);human AC 标
judge: <角色>等 G3。 - 证据会随分支增长而失效:
git diff --stat <首个卡提交>^..HEAD -- <目录>这类开放端点的 原始命令,写下时为真,此后任何非卡的提交落在首个卡提交之后就让它变假——同一条命令,同样的 断言,第二天再跑结果不同(dogfood 2026-09-06,I-79:两条修复提交把"空"变成 5 files / 11 insertions)。 "某类提交没碰某些目录"的承重证据是卡范围的回放仪器(按卡的 commit 列表逐个 replay,如dogfood/ring-audit/replay_card_commits.sh的card_commits_touching_audited_dirs),它只看卡的提交, 分支后续怎么长都不改变结论。原始 diff 只能当补充,且必须钉死两端 + 标记录时刻。 - 体量(按 diff 行数):XS < 50 · S < 200 · M < 500 · L < 1000 · XL ≥ 1000。L 建议拆, XL 必拆(按卡 / 按目录 / 先重构后功能)——long-context 下 review 质量崩,这是经验事实不是偏好。
- push 不可逆、PR 公开:push 之前该确认的都在预检里;PR 建了之后改 body 可以,但已通知的 reviewer 已经读过第一版。
- Draft 的语义:还想再改 / B 档有告警 / 等 CI 结果 → draft;
--draft或预门 flags 非空时默认 draft。 - base 分支:
git config branch.<b>.merge→gh repo view --json defaultBranchRef→ 问用户。 - 关于用户的 Σ:"提上去" = push + 建 PR;"先别通知人" = draft;"顺手合了" ≠ 本 skill 合—— merge 是人类动作。
判据(φ)
- 预检(拒):当前分支 ≠ base;
git status干净;与origin/<base>无落后(有则 merge—— 不 rebase 已推送分支);范围内每条非 merge commit 消息合规(Conventional Commits);自检绿 (入口推断同 /commit)或 body 注明untested。merge 提交的主题由 git 生成,而"落后就 merge"正是 本预检自己的要求——把它按 Conventional Commits 判拒等于照做就过不了(I-74);按 parent 数放行, 非 merge 提交照判不误。 - 插件版本同步(拒):范围 diff 触及
plugins/<p>/skills/**而plugins/<p>/.claude-plugin/plugin.json的 version 未变(或marketplace.json与它不一致)→ 拒(--no-version-sync显式豁免)。 PR 范围看得见整个交付,版本 bump 就该在里面;插件缓存按版本号取目录,不 bump 的修复等于没发。 - body 段落(拒):Summary / Scope(do / dont)/ Linked issue(
Closes #N或Refs #N)/ Changes / Verification(至少一条命令或测试证据)/ Acceptance mapping(有--done-when时每条 mechanical AC 都出现)/ Risk & rollback / Reviewer focus。 - Verification 里的证据形态:命令要么两端钉死(
<sha>^..<记录时的 sha>,并注明记录时刻), 要么用卡范围的回放仪器——以..HEAD收尾的原始 diff 不能承载"某些提交没碰某些目录"这种 否定断言(见 Σ / I-79)。断言在 PR 正文里必须此刻仍为真,不是记录当时为真。 - 体量:XL → 拒(除
--allow-xl并在 body 写明为什么不能拆);L → flag。 - 锁(
--lock或.done_when.lock存在):范围 diff 触及锁文件且无变更提案 → 拒。 - 标题:
type(scope): subject同 commit 规则(≤ 72)。 - 预审(
--pre-review):建 PR 前用新上下文的agents/pr-reviewer.md(跨供应商优先;同源则标 single-vendor caveat)审 base..head 的 diff,最多--rounds(默认 2)轮"修复 → 再审"——业界数据:自审 2–3 轮后收益见顶,且约减少 1/3 的 review 往返。 两轮后仍存活的发现写进 body 的## Known issues(每条带 file:line),不再迭代,留给人看;A 档 / P0 不许作为 known issue 出现——修掉它。verify_pr.py --pre-review检该段存在、每条有锚点、无 A 档。预审不替代 reviewer:它减少的是 nit,不是判断。 - 残差:Summary 是否诚实;Reviewer focus 是否指向真正的风险点。
What ships with it
12 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- assets/pr_template.md 2.0 KB
- eval/fixtures/bad_missing_ac.md 858 B
- eval/fixtures/bad_no_link.md 934 B
- eval/fixtures/bad_no_verification.md 841 B
- eval/fixtures/bad_prereview_noanchor.md 994 B
- eval/fixtures/bad_prereview_p0.md 1022 B
- eval/fixtures/good_body_prereview.md 1.0 KB
- eval/fixtures/good_body.md 931 B
- eval/gate.json 2.1 KB
- eval/report.md 1.7 KB
- references/size-and-split.md 1.6 KB
- scripts/verify_pr.py 16 KB runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- today Changed · +20 lines 711b2d3901b1
- yesterday First seen · 94 lines · 274 tokens per session scan A cac9caafb025
pr is a skill published in the GitHub repository XRenSiu/claude-code-forge (2 stars, last pushed today), licensed MIT. It adds 274 tokens to every session and 2,773 once invoked, about $0.0014 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-05.
Other skills, from other repositories
review-implement-phase
Implements triaged review actions, commits focused fixes, and posts Done plus resolves threads. Use when the user wants only the implementation phase of the review-framework workflow.
github-pr-workflow
Prepare a GitHub pull request from a feature branch — branch hygiene, commit shape, title/body, verification notes, screenshots for UI work, and replies to review comments.
github-automation
GitHub workflow automation, PR management, issue tracking, and code review coordination. Integrates with GitHub Actions and repository management. Use when: PR creation, code review, issue management, release automation, workflow setup. Skip when: local-only changes, non-GitHub repositories.
review-delta
Review only changes since last commit using impact analysis. Token-efficient delta review with automatic blast-radius detection.
work-unit-commits
Plan commits as reviewable work units. Trigger: implementation, commit splitting, chained PRs, or keeping tests and docs with code.
github-contributor
End-to-end playbook for shipping high-quality pull requests to open-source projects you don't maintain — discovery, CONTRIBUTING compliance, PR-size check, minimal-diff implementation, PR description with AI-assisted disclosure, conflict resolution, and post-submission maintainer interaction. Use whenever creating…