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 issuegit 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/issue)<a href="https://agentmods.dev/skills/xrensiu/claude-code-forge/issue"><img src="https://agentmods.dev/badge/skills/xrensiu/claude-code-forge/issue.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 | $0.00283 | $0.02534 |
| Opus 5 | $0.00142 | $0.01267 |
| Sonnet 5 | $0.00057 | $0.00507 |
| Haiku 4.5 | $0.00028 | $0.00253 |
Grade A, and why
issue 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 yesterday.
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 — 115 lines — stays where its author put it; the contents beside it link to each section on GitHub.
issue — 愿望变成能判真假的条目
产物是一条 GitHub issue,其 body 按 assets/issue_template.md 的段落序呈现(这是产物序——
第 5 格——出口可检,不是流程)。本文件写:issue 在世界里是什么、什么算写对、原语与出口、
建之前的门。怎么问、怎么组织材料是你的份额。
缺口(Judgment + Capability + Control)
deletion 测试:撤掉本 skill,引擎会把需求原文当 body 直接建 issue——"让搜索更快""把审核做对"。 这样的 issue 任何实现都能"通过",下游 done_when 没有源头,review 判"越界"没有基准。缺的是 判据(可证伪长什么样)、原语(verify_issue.py 把判据的机械半边编译掉)、门 (建 issue 是公开动作:先给人看)。
世界(Σ)
- issue 是 TASK 的雏形,不是契约本身。 契约(done_when.yaml v2)由
/acceptance-spec或 /sdlc 从这里的 AC 内联生成;issue 里的 AC 已经是 v2 形状,所以两边不会漂移。 - 三种入口,同一模板:
feature(需求)、bug(复现 + 期望 + 实际,AC 就是"复现不再 发生"+ 回归守卫)、escape(合入后发现的缺陷:多一段"归因层 / 为什么该层的门没拦住", 同时追加到escape-defects.md——它是世界层唯一的外部校准源)。 - 两条轨道。 语义份额 × 变化率高、实现原语难度低 → PSL 轨(先
/psl建世界 → G1 人签, 再回来建 issue);确定性主导 → TASK 轨直接建。混合需求只对体验性内核走 PSL。速判:把需求 交给不了解产品的工程师照字面做,会不会做出"技术正确、产品错误"的东西?会 → PSL。 客观触发:--dos dos.yaml时,依赖 DOS:字段里出现 dos.yaml 解析不了的概念 → 强制 PSL 轨(脚本判,不靠自评)。详见references/dual-track.md。 - 范围段四项(做 / 不做 / 硬约束 / 成功度量)是人写的:"做到哪儿为止"agent 推不出来。 你可以起草,但每项都要让用户确认或改。
- 假设台账:澄清最多 3 轮;答不上来的承重槽不默认填,写进台账:假设 · 绑定的 REQ/AC · 风险 · 到期验证点(哪个 AC 或哪道门验证它)。超时不是失败,是带风险签字继续。
- 关于用户的 Σ:"快一点""稳定""大部分"是日常语言,不是判据——把它翻译成数字时,数字
的来源要么是用户给的、要么是 SLO/KPI、要么是一次真实失败;没有来源就
needs_threshold_source标出来问,不编。
判据(φ):什么算写对了
- 可证伪或不进:AC 的 expect / statement 里出现模糊量词(快 / 慢 / 稳定 / 大部分 / 合理 / 友好 / fast / reliable / most / better…)而没有数字阈值 → 不是验收条件。
- happy 带 unhappy 孪生:每条事件型 AC("当 X 时系统应 Y")配一条 unwanted("当 <边界 / 恶意 / 失败输入> 时系统应 <拒绝 / 降级 / 报错>")。未定义的失败语义正是被钻的缝。
- AC v2 形状:
kind: mechanical必有observe(route / cli / ui data-test / db_field / event)given+expect;kind: human必有statement+judge ∈ {product, design, tech}+evidence ∈ {checklist, demo}。详见references/acceptance-shape.md。
- observe 是观察边界,不是实现结构:出现
src/…、.ts/.py/…文件名、函数名 → 拒(契约不该 规定实现结构;文件级存在性下放到任务卡)。 - 范围四项非空;轨道已判定(psl | task);依赖 DOS 字段存在(可为
none,但要写)。 - bug 类:复现步骤 ≥ 1 条、期望 vs 实际都在、至少一条 AC 是回归守卫。
- escape 类:归因层 ∈ {card, plan, task, ontology, world} + "为什么该层的门没拦住"一句。
- 残差(机器不判,路由到人):这些 AC 是不是这次最窄的可证伪条件;阈值是否反映 KPI 而非
空气;unhappy 孪生盖的是不是对的边。脚本给
needs_semantic_review,不裁决。
What ships with it
15 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/bug_extra.md 399 B
- assets/escape_extra.md 442 B
- assets/issue_template.md 2.2 KB
- eval/fixtures/bad_filepath_observe.md 1.7 KB
- eval/fixtures/bad_no_twin.md 1.4 KB
- eval/fixtures/bad_vague.md 1.6 KB
- eval/fixtures/closure_fail.md 1.7 KB
- eval/fixtures/dos.yaml 163 B
- eval/fixtures/good_bug.md 1.8 KB
- eval/fixtures/good_issue.md 1.7 KB
- eval/gate.json 1.7 KB
- eval/report.md 1.7 KB
- references/acceptance-shape.md 2.2 KB
- references/dual-track.md 1.4 KB
- scripts/verify_issue.py 13 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.
- yesterday First seen · 115 lines · 283 tokens per session scan A 0053db4d6b30
issue is a skill published in the GitHub repository XRenSiu/claude-code-forge (2 stars, last pushed yesterday), licensed MIT. It adds 283 tokens to every session and 2,534 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
mcore-split-pr
Split a PR into multiple PRs to reduce the number of required CODEOWNERS reviewer groups.
shiplog
Recap of everything shipped since the last run - cross-repo PRs, security fixes, star deltas, and X traction, synthesized into a digest article and a ready-to-post shiplog in your voice.
llxprt-issue-workflow
Use this skill when asked to address, fix, or work on a GitHub issue in the llxprt-code repository. Covers the complete issue lifecycle - branch setup, gh CLI usage, test-first planning, subagent delegation and review loops, the full verification cycle (including the stepfun-37 smoke test), open code review (ocr), PR…
cyrus-setup-repository
Add one or more Git repositories to Cyrus configuration so it can process issues from those repos.
github-sync
Bidirectional synchronization of epics and tasks with GitHub issues, labels, and relationships.
github-sync-helper
General GitHub basic operations + automation for GitHub platform objects (Issues/Labels/Milestones/Releases/Actions) in the Minis environment. This skill must be triggered when the user mentions any basic Git/GitHub operation or workflow, including "how to use GitHub," clone, init, remote, branch, commit, push, pull…