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 tunegit 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/tune)<a href="https://agentmods.dev/skills/xrensiu/claude-code-forge/tune"><img src="https://agentmods.dev/badge/skills/xrensiu/claude-code-forge/tune/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/xrensiu/claude-code-forge/tune"><img src="https://agentmods.dev/badge/skills/xrensiu/claude-code-forge/tune.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00339 | $0.02745 |
| Opus 5 | $0.00169 | $0.01373 |
| Sonnet 5 | $0.00068 | $0.00549 |
| Haiku 4.5 | $0.00034 | $0.00275 |
Grade A, and why
tune 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 3d ago.
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 — 117 lines — stays where its author put it; the contents beside it link to each section on GitHub.
tune — 环改环的外环，改动永不自动生效
产物：tune/harness-proposals-<date>.yaml（形状 assets/harness_proposals_template.yaml）与可选的 tune/*.patch。
本文件写：这个环在流水线里是什么、什么算一条合格的 harness 提案、原语与出口、不可豁免的约束。哪条提案值得开 PR 是人的份额。
缺口（Knowledge + Judgment + Control）
deletion 测试：撤掉本 skill，引擎会在某次 review 跑到 exit 30 后说"MAX_ROUNDS 太小了，改成 20"——没有跨 PR 的数据、 没有对照"是预算不够还是层错了"、直接 sed 改脚本。缺的是：数据在哪、怎么算（Σ）；什么算 harness 参数需要动的信号， 什么只是一次噪声（φ）；导出与出 diff 的脚本（Π）；提案永不自动生效的门（γ）。
世界（Σ）
-
这是 hill_climb 环（
../ai-dlc/assets/loops.yaml#hill_climb）：generator = 本 skill，verifier =human.harness-review， timescale = days，trigger = cron。它读其他五个环的 memory 文件，不读实现者会话。 -
数据全部来自已有产物：
specs/<slug>/{state.json, ledger.md, trace.jsonl, escape-defects.md}、.aidlc/pr-watch/pr-N.json+pr-N.counters.json、ratchet-log/**/results.tsv；参数现值来自routing.yaml/loops.yaml/pr-poll.sh的默认值。不需要额外埋点。 -
可动的东西是封闭集（target ∈）：
routing.budgets.<track>.<key>·routing.fingerprint_repeat_limit·routing.plateau_rounds·review-loop.MAX_ROUNDS·review-loop.MAX_THREAD_STRIKES·acceptance-fleet.isolation_min·pr-review.b_tier_thresholds·code-reviewer.focus_allocation·<skill>.fix_list。契约、世界、代码、skill 正文都不在集合里——它们分别是 /retro、/implement、 skill-evolve 的地盘。 -
信号与它说明的问题：
统计 说明 某层预算从未被碰（max_used = 0） 预算可能过宽——降 1，不影响任何已发生的升级 某层预算在 ≥ 50% 特性上耗尽 且 没有 repeat / oscillation / plateau 升级 重试在推进，预算真的短——升 1 预算耗尽 且伴随 收敛升级 不是预算问题，是层错了——不提案，写 note review 轮次最大值 ≤ 40% MAX_ROUNDS 且从未 exit 30 MAX_ROUNDS 可收紧（失控环更早停） ≥ 50% PR 撞到 MAX_ROUNDS 升 2，但风险是掩盖本该去 G3 的争议 ACCEPT / 已裁决 ≥ 0.95 且线程 ≥ 5 sycophancy 反向代理（SWE-Review）：REJECT 是合法且被期望的；进 review-loop 的 fix_list strike 冻结线程 ≥ 50% PR 先审 REJECT 回帖有没有证据，不动 MAX_THREAD_STRIKES 逃逸缺陷按归因层 进对应层 skill 的 fix_list（task → donewhen-extract，plan → plan-cards，card → implement …）； trace.py why看哪道门漏的豁免出现在 ≥ 50% 特性 前置条件可能定错了，不是人不小心——进 AI-DLC 的 fix_list G1 reject 以 rule_error 为主 PSL 在推导分歧处欠定——进 psl 的 fix_list
What ships with it
19 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/harness_proposals_template.yaml 2.8 KB
- eval/fixtures/archive-single/feat-a/done_when.yaml 1.4 KB
- eval/fixtures/archive-single/feat-a/escape-defects.md 201 B
- eval/fixtures/archive-single/feat-a/state.json 573 B
- eval/fixtures/archive-single/feat-a/trace.jsonl 3.2 KB
- eval/fixtures/archive/feat-a/done_when.yaml 1.4 KB
- eval/fixtures/archive/feat-a/escape-defects.md 201 B
- eval/fixtures/archive/feat-a/state.json 573 B
- eval/fixtures/archive/feat-a/trace.jsonl 3.2 KB
- eval/fixtures/archive/feat-b/state.json 730 B
- eval/fixtures/archive/feat-b/trace.jsonl 1.9 KB
- eval/fixtures/pr-watch/pr-12.counters.json 44 B
- eval/fixtures/pr-watch/pr-12.json 1.0 KB
- eval/fixtures/pr-watch/pr-15.counters.json 54 B
- eval/fixtures/pr-watch/pr-15.json 519 B
- eval/gate.json 1.8 KB
- eval/report.md 885 B
- scripts/apply_proposal.py 6.1 KB runs code
- scripts/tune.py 23 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.
- 3d ago Changed da586c60f270
- 5d ago First seen · 117 lines · 339 tokens per session scan A ad2f686a8838
tune is a skill published in the GitHub repository XRenSiu/claude-code-forge (2 stars, last pushed today), licensed MIT. It adds 339 tokens to every session and 2,745 once invoked, about $0.0017 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 2026-09-07.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
insight-error-page
Write or audit an insight-kind error page for the Next.js dev overlay. Use when creating a new errors/ .mdx page, auditing an existing one, or checking that a page matches the framework fix cards. Covers page structure, title alignment, FixCard cards with Copy prompt button, code snippets, terminology verification…