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 scandnavik/writing-harness --skill tightengit clone --depth 1 https://github.com/scandnavik/writing-harnessWrote 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/scandnavik/writing-harness/tighten)<a href="https://agentmods.dev/skills/scandnavik/writing-harness/tighten"><img src="https://agentmods.dev/badge/skills/scandnavik/writing-harness/tighten/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/scandnavik/writing-harness/tighten"><img src="https://agentmods.dev/badge/skills/scandnavik/writing-harness/tighten.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 | $0.00086 | $0.01764 |
| Opus 5 | $0.00043 | $0.00882 |
| Sonnet 5 | $0.00017 | $0.00353 |
| Haiku 4.5 | $0.00009 | $0.00176 |
Grade A, and why
tighten 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 10d 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 — 138 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Tighten
Markdown 精簡管線，目的是省 token。分工是這樣：偵測交給腳本（不花 LLM），只有重寫才動用 LLM，避免把便宜的偵測工作浪費在昂貴的模型上。
bloat（填充詞、無用前綴、重複框架、description 回聲）是 AI slop 的另一張臉：不是寫錯，是寫太多、講太繞。這支 skill 把「偵測」交給純 regex（零 LLM），只把「重寫」交給 LLM，最省 token。
預設行為
| 輸入 | 預設掃描路徑 |
|---|---|
| 無參數 | **/*.md（排除你指定的 index／生成檔） |
| 單一路徑 | 該檔或該 glob |
| 多路徑 | 全部納入 |
建議絕不碰：結構敏感的索引檔（手工維護的目錄）、機器生成的 yaml、工具腳本目錄。把這些路徑加進你自己的排除清單。
10 條規則（詳見 scripts/verbosity-check.py）
| # | 規則 | 一句話 |
|---|---|---|
| 1 | frontmatter-description-echo | description 重複 name 字詞 >60% |
| 2 | dated-parenthetical | body 中「（YYYY-MM-DD 補）」冗贅 |
| 3 | marker-prefix | 結論 / 教訓 / 策略 等無用前綴 |
| 4 | meta-blockquote | H1 後首段 > 分類原則/說明/規則 |
| 5 | numbered-code-comments | code block ≥3 個 # N. 編號註解 |
| 6 | filler-words | 其實/當然/基本上/事實上... |
| 7 | dual-preamble | 情境+Pattern 或 方案+做法 |
| 8 | h1-h2-echo | H1 後立刻 H2 同名 |
| 9 | list-prefix-bloat | list 以「這是/這段是/這個方案/這個功能」開頭 |
| 10 | dup-pros-apply | 同節 優點 + 適用 雙 bullet 列 |
執行流程
Step 1 — 執行偵測腳本
python ~/.claude/skills/writing-harness/scripts/verbosity-check.py <paths> --format=markdown
讀 exit code：
0→ 回報「檔案已足夠精簡」，退出1→ 繼續 Step 22→ 回報錯誤，退出
Step 2 — 呈現 report 並問核准
把 markdown report 貼給用戶，附一行摘要：
共 N 處命中、散佈在 M 個檔案。各規則命中次數：<rule>: <count>, ...
要重寫嗎？(y / n / 指定檔案子集)
等待用戶回應。
Step 3 — 派 sonnet sub-agent 重寫
對每一個有 findings 的檔案（依用戶選的子集），spawn 一個 general-purpose agent（model: sonnet）。
Agent prompt 模板：
你是精簡寫手。下方 .md 檔的特定段落被標記為冗贅（偵測規則、行號、snippet）。
任務：只修 flagged 段落，其餘原封不動。
硬性約束：
1. 不動 frontmatter 的 name / type / 任何系統欄位
2. frontmatter 的 description 可改（若被 flagged 為 description-echo）
3. 保留所有 markdown links `[text](path)`、code blocks、table 結構
4. 不新增段落、不改 H1 標題
5. 語意等價：每條資訊都要留，只砍表達
6. 不確定時 → 保留原文（寧願 under-cut 不要 over-cut）
規則對照：
- marker-prefix：刪掉「**結論**：」「**教訓**：」等前綴，保留後面的內容
- dated-parenthetical：刪掉「（YYYY-MM-DD 補）」這類標記
- filler-words：刪「其實/當然/基本上」等空泛修飾
- meta-blockquote：刪掉整行 `> 分類原則：...`
- dual-preamble：兩段合併為一句開場
- h1-h2-echo：刪掉重複的 H2 或改成細節標題
- list-prefix-bloat：刪「- 這是 / - 這段是」開頭的冗贅
- numbered-code-comments：code 內 `# 1. # 2. # 3.` 保留 1-2 個關鍵註解即可
- dup-pros-apply：flag 即可，若語意真的重疊才合併；否則保留
- frontmatter-description-echo：重寫 description 講「這份文件提供什麼」（不重複 name）
輸出：直接用 Edit/Write 改檔，不要回文字報告。改完後用一句話說「<file> 已精簡」。
檔案路徑：<path>
Findings：
<list of findings with line + rule + snippet>
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.
- 10d ago First seen · 138 lines · 86 tokens per session scan A 5ac14aeb58ff
tighten is a skill published in the GitHub repository scandnavik/writing-harness (23 stars, last pushed 2mo ago), licensed MIT. It adds 86 tokens to every session and 1,764 once invoked, about $0.0004 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 2026-08-30.
Other skills, from other repositories
writing
Draft, revise, audit, or transform human-facing prose: articles, blogs, documentation, criticism, essays, email, marketing and SEO copy, summaries, scripts, application materials, and UI text. Excludes code comments, commit messages, and private notes.
happiness-skill
A Chinese-language guide to happiness based on reducing unmet wants, focusing on the present, and treating happiness as a trainable skill.
setup-matt-pocock-skills
A setup skill that configures engineering skills for a repository, including its issue tracker, labels, and documentation layout. A repository is the project folder managed by version control.
frontend-design
A design guide for building polished web interfaces such as pages, dashboards, forms, navigation, and reusable UI components. It covers HTML, CSS, JavaScript, and common frontend frameworks.
alterlab-cobrapy
Build and analyze genome-scale constraint-based metabolic models with COBRApy — flux balance analysis (FBA), flux variability analysis (FVA), gene and reaction knockouts, flux sampling, and SBML model I/O. Use when simulating metabolic networks, predicting growth or knockout phenotypes, or running systems-biology and…
alterlab-depmap
Query the Cancer Dependency Map (DepMap) for cancer cell line gene dependency scores (CRISPR Chronos), drug sensitivity data, and gene effect profiles. Use when identifying cancer-specific genetic vulnerabilities, finding synthetic lethal interactions, checking whether a gene is essential in given cell lines, or…