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 agentmods add skills/arcadi4/nerdy/approximation-algorithmsnpx skills add Arcadi4/nerdy --skill approximation-algorithmsgit clone --depth 1 https://github.com/Arcadi4/nerdyWhat 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 | $0.00075 | $0.04400 |
| Opus 5 | $0.00037 | $0.02200 |
| Sonnet 5 | $0.00015 | $0.00880 |
| Haiku 4.5 | $0.00007 | $0.00440 |
Grade A, and why
approximation-algorithms 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 2d 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 — 422 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Approximation Algorithms
Overview
Approximation answers are certificate audits, not heuristic endorsements. First decide whether the problem is minimization or maximization, identify a polynomial-time lower or upper bound on the optimum, then prove the returned feasible solution is within the promised factor under the chapter's exact preconditions.
Core principle: never claim an approximation ratio from a good-looking local rule. Name the optimum bound, prove feasibility, prove the ratio against that bound, and reject the claim if any precondition, encoding assumption, or shortcut inequality is missing.
Shared CLRS Conventions
- Follow the parent
clrsskill for mathematical formatting: every cost ratio, probability, recurrence, inequality, asymptotic bound, threshold, and approximation factor belongs in a display LaTeX block. - Keep tables verbal. Put ratio definitions, LP constraints, probabilities, trimming guarantees, and running times in display blocks near the table instead of inside table cells.
- Use
np-completenessfor exact decision hardness and reductions; use this skill when the task asks for near-optimal algorithms, approximation ratios, inapproximability gaps, schemes, or approximation proof certificates. - Use
linear-programmingfor general LP modeling; use this skill when an LP relaxation is being rounded to certify an approximation algorithm. - Do not announce that you are using this skill. Deliver the polished algorithm, proof, counterexample, or review directly.
When to Use
Use this skill for:
- proving or refuting an approximation ratio for a polynomial-time algorithm;
- distinguishing exact, pseudo-polynomial, PTAS, and FPTAS claims;
- checking metric TSP, general TSP, vertex cover, set cover, MAX-3-CNF, weighted vertex cover, or subset-sum approximation arguments;
- turning a greedy, randomized, LP-rounding, or trimming proof into a certificate-style answer;
- reviewing chapter problem patterns such as bin packing, weighted set cover, maximal matching, list scheduling, maximum spanning tree, maximum clique, or knapsack approximations.
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.
- 2d ago First seen · 422 lines · 75 tokens per session scan A 891646268ec4
approximation-algorithms is a skill published in the GitHub repository Arcadi4/nerdy (7 stars, last pushed 4mo ago), licensed MIT. It adds 75 tokens to every session and 4,400 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 2026-08-31.
Other skills, from other repositories
pptx
从论文、大纲或结构化文本生成 PowerPoint (.pptx) 演示文稿。Use when 用户需要把一篇论文/文章/大纲做成幻灯片、slides、演示文稿、PPT、deck。Don't use when 只需纯文本总结、生成 Word/PDF、或修改已有 pptx 的单个像素级样式。.
ai-style
当任务是用中文撰写或改写面向读者的文案（产品发布稿、公众号文章、邮件、README 等）， 或用户反馈文字「AI 味太重」「不像人写的」时，加载本 Skill。.
curly-quote-sft
Skill "curly-quote-sft" from bojieli/ai-agent-book, covering 中文技术文档符号与引用规范, 何时加载, 符号定义, 决策优先级 and 正反例约束.
triage
你是当前任务的分诊协调者。先识别用户的全部目标、顺序依赖和验收条件，再按 “事实检索 → 计算/执行 → 写作”顺序逐步请求切换到需要的专业能力。不要替专业 能力完成它的工作，也不要在信息缺失时臆造结果。.
writing
将共享历史中的已验证事实和计算结果整理成符合受众、格式与长度约束的成稿。.
research
用真实检索工具查找可追溯的事实、数据和来源。.