Agent skill

Add Benchmark

為 Twinkle Eval 新增一個評測 benchmark（IFEval、BFCL、RAGAS、Text2SQL、Vision MCQ 之類）。涵蓋 CLAUDE.md §6 的完整強制流程：先建 Milestone 與 6 個 Issue、準備 example dataset、實作 Extractor + Scorer 並註冊 PRESETS、與參考框架做分數與速度對比、撰寫…

MITAuto-check passedAI & LLM Engineering

Install Add Benchmark

skills CLI

$ npx skills add ai-twinkle/Eval --skill add-benchmark -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI

$ gh skill install ai-twinkle/Eval add-benchmark --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy

$ git clone --depth 1 https://github.com/ai-twinkle/Eval.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.claude/skills/add-benchmark .claude/skills/add-benchmark && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Paste into Claude Code

Install the "add-benchmark" agent skill from https://github.com/ai-twinkle/Eval/tree/main/.claude/skills/add-benchmark into .claude/skills/add-benchmark/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "add-benchmark", then confirm the skill loads.

Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

$skill-installer

$skill-installer install https://github.com/ai-twinkle/Eval/tree/main/.claude/skills/add-benchmark

Type this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.

skills CLI

$ npx skills add ai-twinkle/Eval --skill add-benchmark -a codex

Project install goes to .agents/skills/; add -g for ~/.codex/skills/.

GitHub CLI

$ gh skill install ai-twinkle/Eval add-benchmark --agent codex

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/ai-twinkle/Eval.git skills-src && mkdir -p .agents/skills && cp -r skills-src/.claude/skills/add-benchmark .agents/skills/add-benchmark && rm -rf skills-src

Use ~/.agents/skills/ instead of .agents/skills for a personal install.

Codex skills documentation · loads skills from .agents/skills/

Paste into Codex

Install the "add-benchmark" agent skill from https://github.com/ai-twinkle/Eval/tree/main/.claude/skills/add-benchmark into .agents/skills/add-benchmark/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "add-benchmark", then confirm the skill loads.

Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

skills CLI

$ npx skills add ai-twinkle/Eval --skill add-benchmark -a cursor

Project install goes to .agents/skills/; add -g for ~/.cursor/skills/.

GitHub CLI

$ gh skill install ai-twinkle/Eval add-benchmark --agent cursor

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/ai-twinkle/Eval.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/.claude/skills/add-benchmark .cursor/skills/add-benchmark && rm -rf skills-src

Use ~/.cursor/skills/ instead of .cursor/skills for a personal install.

Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/

Paste into Cursor

Install the "add-benchmark" agent skill from https://github.com/ai-twinkle/Eval/tree/main/.claude/skills/add-benchmark into .cursor/skills/add-benchmark/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "add-benchmark", then confirm the skill loads.

Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

gemini skills install

$ gemini skills install https://github.com/ai-twinkle/Eval.git --path .claude/skills/add-benchmark

--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.

skills CLI

$ npx skills add ai-twinkle/Eval --skill add-benchmark -a gemini-cli

Project install goes to .agents/skills/; add -g for ~/.gemini/skills/.

GitHub CLI

$ gh skill install ai-twinkle/Eval add-benchmark --agent gemini-cli

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/ai-twinkle/Eval.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/.claude/skills/add-benchmark .gemini/skills/add-benchmark && rm -rf skills-src

Use ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.

Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/

Paste into Gemini CLI

Install the "add-benchmark" agent skill from https://github.com/ai-twinkle/Eval/tree/main/.claude/skills/add-benchmark into .gemini/skills/add-benchmark/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "add-benchmark", then confirm the skill loads.

Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

GitHub CLI

$ gh skill install ai-twinkle/Eval add-benchmark

Installs for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).

skills CLI

$ npx skills add ai-twinkle/Eval --skill add-benchmark -a github-copilot

Project install goes to .agents/skills/; add -g for ~/.copilot/skills/.

Manual copy

$ git clone --depth 1 https://github.com/ai-twinkle/Eval.git skills-src && mkdir -p .github/skills && cp -r skills-src/.claude/skills/add-benchmark .github/skills/add-benchmark && rm -rf skills-src

Use ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.

GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/

Paste into GitHub Copilot

Install the "add-benchmark" agent skill from https://github.com/ai-twinkle/Eval/tree/main/.claude/skills/add-benchmark into .github/skills/add-benchmark/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "add-benchmark", then confirm the skill loads.

GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

skills CLI

$ npx skills add ai-twinkle/Eval --skill add-benchmark -a opencode

OpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.

GitHub CLI

$ gh skill install ai-twinkle/Eval add-benchmark --agent opencode

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/ai-twinkle/Eval.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/.claude/skills/add-benchmark .opencode/skills/add-benchmark && rm -rf skills-src

Use ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.

OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/

Paste into OpenCode

Install the "add-benchmark" agent skill from https://github.com/ai-twinkle/Eval/tree/main/.claude/skills/add-benchmark into .opencode/skills/add-benchmark/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "add-benchmark", then confirm the skill loads.

OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

Facts

Skill name: add-benchmark
GitHub stars: 117
Token cost: ~1.8k tokens
SKILL.md length: 501 words
Files: 1
Skills in repo: 2
Repo updated: Sep 15, 2026
First seen: Oct 7, 2026
Licence: MIT

At a glance

Works in 3 steps: 「資料集清單」表格新增條目（目錄、來源、題數、評測方法、說明） → 「快速開始」新增對應的 config.yaml 範例 → 「資料格式」新增該 benchmark 的 JSONL 格式說明
Tasks that involve LLM evaluation
SKILL.md covers 動手前先確認, 第 0 步：先開 GitHub，再寫程式, 第 1 步：example dataset and 第 2 步：實作 Extractor + Scorer, plus 6 more sections
Calls gh and python3

What it does

Add Benchmark is an agent skill from ai-twinkle/Eval. 為 Twinkle Eval 新增一個評測 benchmark（IFEval、BFCL、RAGAS、Text2SQL、Vision MCQ 之類）。涵蓋 CLAUDE.md §6 的完整強制流程：先建 Milestone 與 6 個 Issue、準備 example dataset、實作 Extractor + Scorer 並註冊 PRESETS、與參考框架做分數與速度對比、撰寫 docs/evals/{name}.md、更新 datasets/example/README.md、建立 tests/test{name}.py。當使用者說「新增 benchmark」「加一個評測方法」「支援 XXX 評測」「實作 evaluationmethod」時使用。

Its SKILL.md is about 1.8k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in AI & LLM Engineering, covering LLM evaluation, Project management and Agent instruction files. The repository describes itself as: High-performance LLM evaluation framework with parallel API calls — up to 17× faster than sequential tools. Supports box, math, and logit-based evaluation. The licence is MIT.

When your agent uses it

Tasks that involve LLM evaluation
Tasks that involve Project management
Tasks that involve Agent instruction files

Example prompts

“/add-benchmark”

Requirements

Python 3

Workflow steps

3 steps, taken from the first numbered list in SKILL.md.

「資料集清單」表格新增條目（目錄、來源、題數、評測方法、說明）
「快速開始」新增對應的 config.yaml 範例
「資料格式」新增該 benchmark 的 JSONL 格式說明

What it can do on your machine

Read from SKILL.md and the folder at commit 608273c. It shows what the files ask for, not the result of running them.

Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
Shell commands in SKILL.md call:
- gh
- python3
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md. Its commands use gh, which can reach the network depending on how they are called.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Add Benchmark loads about 1.8k tokens when it runs. Until then it costs about 87 tokens; SKILL.md has 501 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it: ~87
When it runs · the whole SKILL.md, loaded when a task matches: ~1.8k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from ai-twinkle/Eval at commit 608273c, republished under its MIT licence (© ai-twinkle). 501 words, ~1,775 tokens.

Download SKILL.mdSave it as .claude/skills/add-benchmark/SKILL.md (or your agent's skills folder).

name: add-benchmark
description: 為 Twinkle Eval 新增一個評測 benchmark（IFEval、BFCL、RAGAS、Text2SQL、Vision MCQ 之類）。涵蓋 CLAUDE.md §6 的完整強制流程：先建 Milestone 與 6 個 Issue、準備 example dataset、實作 Extractor + Scorer 並註冊 PRESETS、與參考框架做分數與速度對比、撰寫 docs/evals/{name}.md、更新 datasets/example/README.md、建立 tests/test_{name}.py。當使用者說「新增 benchmark」「加一個評測方法」「支援 XXX 評測」「實作 evaluation_method」時使用。

新增評測 Benchmark

CLAUDE.md §6 規定的流程，缺一步就不得合入 main。這份 skill 是那份規範的可執行版本。

動手前先確認

這是新的評測方法（新的 Extractor/Scorer 配對），不是既有方法的新資料集。若只是新資料集且能沿用既有 evaluation_method，走 twinkle_eval/benchmarks.py 的下載註冊表即可，不需要跑本流程。
上一個 benchmark 已經完成（含比較報告）。§6.0.1 明訂：先完成再開始下一個。

第 0 步：先開 GitHub，再寫程式

沒有 Milestone 和 Issues 的 benchmark 實作不得開 PR。

bash

gh api repos/ai-twinkle/Eval/milestones -f title="{Benchmark} — {一句話說明}" \
  -f description="來源：{paper/repo}｜目的：{評什麼}｜範圍：{預計實作到哪}"

接著開 6 個 Issue，每個都要掛 type: feature label 與該 Milestone：

#	標題格式	對應
1	`feat({name}): prepare example dataset from HuggingFace`	§6.1
2	`feat({name}): implement Extractor + Scorer (+ Checkers)`	§6.2
3	`feat({name}): score comparison vs reference framework`	§6.3
4	`feat({name}): speed benchmark vs reference framework`	§6.4
5	`feat({name}): write docs/evals/{name}.md`	§6.5
6	`feat({name}): write tests/test_{name}.py`	§6.6

bash

gh issue create --title "feat({name}): ..." --label "type: feature" --milestone "{Milestone 標題}"

第 1 步：example dataset

放在 datasets/example/{name}/，10–20 筆（§6.1 規範；既有資料集實際落在 10–30 筆，如 aime2025 收了全部 30 題），涵蓋主要題型分佈。有子類別的（如 BFCL 的 simple/multiple/parallel）每個子類別至少 2–3 筆。

要求任何人不下載完整資料集就能跑通完整流程。格式必須是可直接執行的完整格式（含 id、question、答案欄位）。

scripts/ 底下有既有的抽取腳本可以參考（prepare_bbh_example.py、 create_vision_mcq_example.py）。

資料集載入器支援 JSON / JSONL / CSV / TSV / Parquet / Arrow。twinkle_eval/datasets/file.py 的 _normalize_record() 會自動把 {"choices": [...], "answer": 1} 這種 HuggingFace 格式正規化成 {"A": ..., "B": ..., "answer": "B"}。

第 2 步：實作 Extractor + Scorer

架構是 Extractor（從輸出抽答案）+ Scorer（比對正解） 兩個 ABC，定義在 twinkle_eval/core/abc.py。不要在 evaluators.py / main.py 加 if/elif 分支（原則 A）。

python

# twinkle_eval/metrics/extractors/{name}.py
from typing import Optional

from twinkle_eval.core.abc import Extractor


class MyExtractor(Extractor):
    def get_name(self) -> str:
        return "{name}"

    def extract(self, llm_output: str) -> Optional[str]:
        """抽不到回傳 None（會被計入 unparsed_count）。"""
        ...

python

# twinkle_eval/metrics/scorers/{name}.py
from twinkle_eval.core.abc import Scorer


class MyScorer(Scorer):
    def get_name(self) -> str:
        return "{name}"

    def normalize(self, answer: str) -> str:
        return answer.strip().upper()

    def score(self, predicted: str, gold: str) -> bool:
        ...

註冊到 twinkle_eval/metrics/__init__.py 的 PRESETS：

python

PRESETS: Dict[str, Tuple[Type[Extractor], Type[Scorer]]] = {
    ...
    "{name}": (MyExtractor, MyScorer),
}

同時把類別加進該檔案的 __all__，並視需要加進 twinkle_eval/__init__.py。

Evaluator 需要新的資料流嗎？

runners/evaluator.py 的 evaluate_file() 依 extractor 上的 flag 分流。若你的 benchmark 不是「送文字、收文字」，設對應的 class attribute：

Flag	路徑	既有使用者
`uses_logprobs`	逐選項算 log-likelihood，不生成	logit
`uses_tool_calls`	送 `tools`，讀 `message.tool_calls`	bfcl_fc
`uses_prompt_injection`	把 function 定義注入 system prompt	bfcl_prompt
`uses_ifeval`	傳 `instruction_id_list` + `kwargs` 給 checker	ifeval, ifbench
`uses_audio`	送音檔（Whisper API 或多模態）	asr
`uses_vision`	送圖片（base64 data URI 或 URL）	vision_mcq
（無）	純文字解析	pattern, box, math, ...

新增第 8 條路徑等於「在核心流程新增大量邏輯」，依 §7 必須先開 Issue 取得 maintainer 同意。先確認能不能用既有路徑。

需要多指標（不只 accuracy）？

在 Scorer 上實作 score_full() 回傳 dict（IFEval 的四個指標、ASR 的 WER/CER 都是這樣做）。

⚠️ 但 evaluator 目前只在 uses_ifeval 與 uses_audio 兩條路徑呼叫 score_full()。 文字、logit、tool_calls、prompt_injection、vision 五條路徑不會呼叫，在那些路徑上實作了也不會生效。若新 benchmark 需要多指標又不走這兩條路徑，得先在 evaluator 加上呼叫點——那屬於「在核心流程新增邏輯」，依 §7 要先開 Issue。

移植第三方程式碼

不得整份複製參考框架，只萃取核心評分邏輯
Python 檔案頂部加 attribution 註解，docs/evals/{name}.md 記錄授權（Apache 2.0 / MIT）
原始碼用到不可用的依賴（absl、immutabledict）必須換成標準庫等價物
新依賴一律放 pyproject.toml 的 [project.optional-dependencies]。新增 required dependency 依 §7 必須先開 Issue 並取得 maintainer 同意，且不得讓單機執行路徑增加必要依賴（原則 H）

第 3 步：config 範本

twinkle_eval/templates/{name}.yaml。--init 會自動掃描這個目錄，不需要改程式碼。

範本必須含 llm_api / model / evaluation / logging 四個區塊，並在註解裡寫清楚 optional deps 的安裝指令與資料集欄位需求。參考 templates/ifeval.yaml。

若有 strategy_config 參數，在範本裡列出全部並附預設值。

Show full SKILL.md (203 more words)Show less

第 4 步：分數對比（§6.3）

用相同模型、相同題目同時跑本專案與參考框架。

資料集大小	可接受誤差
≥ 200 筆	±2%
50–199 筆	±3%
< 50 筆	±5%
≤ 20 筆（example）	僅 sanity check，不強制對比

超出容差必須查明原因（通常是 preprocessing、prompt 格式、或答案正規化邏輯差異）並在文件說明。

第 5 步：速度對比（§6.4）

速度是本專案的核心賣點，必須記錄：本專案總耗時 / 並行 worker 數 / 模型名稱，以及參考框架在同等硬體、同等題數下的耗時。無法直接對比就註明原因。

第 6 步：文件

bash

cp docs/evals/TEMPLATE.md docs/evals/{name}.md

填完所有 {...} 欄位，特別是第 6 節（分數對比）與第 7 節（速度對比）——這兩節空著不得合入。

同時更新 datasets/example/README.md（§6.5.1）

三個地方都要改，漏一個就不符合合入條件：

「資料集清單」表格新增條目（目錄、來源、題數、評測方法、說明）
「快速開始」新增對應的 config.yaml 範例
「資料格式」新增該 benchmark 的 JSONL 格式說明

第 7 步：測試（§6.6）

tests/test_{name}.py，必須能在不呼叫任何外部 API 的情況下通過（pure unit test）。必備覆蓋：

Extractor — get_name()、extract() 行為、flag（uses_*）
Scorer — get_name()、normalize()、score() 的正確/錯誤/空值/無效 ground truth
score_full()（若有）— 回傳結構驗證
Checker / 評分邏輯 — 挑 3–5 種代表性 rule type，各寫一個 pass 與一個 fail
Checker Registry（若有）— 所有指令 ID 都已註冊、數量正確、category 齊全
PRESETS 註冊 — PRESETS["{name}"] 存在且對應正確的類別
Example Dataset — 檔案存在、必要欄位齊全、筆數符合、涵蓋所有子類別
Edge cases — 空回應、None、kwargs 內的 null 過濾

bash

python3 -m pytest tests/test_{name}.py -v    # 新增測試全過
python3 -m pytest tests/ -v                  # 完整套件無新增失敗

既有失敗（版本不一致、缺本機 fixture）要確認在本次變更之前就存在，並在 PR 描述說明。

第 8 步：合入前檢查

§6.0.1 的合入前提，全部要打勾：

6 個 Issue 都已 close（或在同一 PR 解決）
docs/evals/{name}.md 完整，含分數對比與速度對比
分數誤差符合容差標準
datasets/example/{name}/ 有 10–20 筆
datasets/example/README.md 三處都更新
tests/test_{name}.py 全過，完整套件無新增失敗
移植的程式碼有 attribution 註解與授權記錄
新依賴都在 [project.optional-dependencies]

push 前：強制 reviewer agent（§13）

這條沒有例外。 開 PR 前、以及後續每一次 push 新 commit 前，都必須 spawn 一個獨立的 reviewer agent 審自己的 diff：

Agent(subagent_type="general-purpose", prompt="""
審查 {branch} 的變更。先讀 CLAUDE.md，再讀 git diff main...HEAD。
變更動機：{為什麼這樣改}
找：bug、違反 CLAUDE.md 原則之處、邊界條件、測試漏洞。
回報 BLOCKER / SHOULD-FIX / NIT 三層，每項附檔案:行號與具體觸發情境。
""")

Reviewer 回報 blocker 必須在 push 前修掉
should-fix / nit 至少要在 PR 描述說明處置方式（修了 / 開 follow-up / 為何不修）
不得自己 review 自己——coding agent 對剛寫的程式碼有確認偏差

版號（§11）

Milestone 全部 Issue close → bump MINOR（新增 benchmark 屬新功能）。同步改 pyproject.toml 與 twinkle_eval/__init__.py，更新 CHANGELOG.md，建 tag 並 gh release create，然後 close Milestone。

Files

Just SKILL.md in .claude/skills/add-benchmark of ai-twinkle/Eval.

Open the folder on GitHubat commit 608273c

Compare with similar skills

Add Benchmark next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Add Benchmark compared with similar skills
Skill	Stars	Used in	Tokens	Auto-check	Licence	Repo updated
Add Benchmark this skillai-twinkle/Eval	117	—	~1.8k	Automated safety check: Pass	MIT	23 days ago
Context Auditundefined-ui/second-brain-os	1k	—	~802	Automated safety check: Pass	MIT	yesterday
Docs SyncKiln-AI/Kiln	5.2k	—	~1.7k	Automated safety check: Pass	Custom licence	yesterday
Create Simple Promptpnp/copilot-prompts	892	—	~2.6k	Automated safety check: Pass	MIT	2 days ago
Write Skilldruxt/druxt.js	114	—	~926	Automated safety check: Pass	MIT	today
Neat-Freak Knowledge CloseoutKKKKhazix/khazix-skills	21k	—	~1.9k	Automated safety check: Pass	MIT	7 days ago

Similar skills

Context Audit
undefined-ui/second-brain-os
Audit an agent's context layout against the four places: system prompt, tools, history, tail.
1k GitHub stars~802 tokensUpdated yesterday
AI & LLM EngineeringAuto-check passed
Docs Sync
Kiln-AI/Kiln
Find and fix the docs and agent guidance that a branch makes stale - agent skills and their reference files, AGENTS.md and .agents/.md prompts, READMEs, docs/ folders, and code-adjacent docs that…
5.2k GitHub stars~1.7k tokensUpdated yesterday
AI & LLM EngineeringAuto-check passed
Create Simple Prompt
pnp/copilot-prompts
This skill should be used when the user asks to "create a new prompt sample", "add a new prompt sample", "scaffold a new prompt sample", "create a prompt contribution", "add a prompt", or needs to…
892 GitHub stars~2.6k tokensUpdated 2 days ago
AI & LLM EngineeringAuto-check passed
Write Skill
druxt/druxt.js
Creates or changes a druxt.js contributor skill in .agents/skills, with its evals and the tests that gate it.
114 GitHub stars~926 tokensUpdated today
Agent WorkflowsAuto-check passed
Neat-Freak Knowledge Closeout
KKKKhazix/khazix-skills
Brings project docs, agent rule files, authorized memory and leftover workspace files back in line with what the code and runtime actually do at the end of a work session.
21k GitHub stars~1.9k tokensUpdated 7 days ago
Agent WorkflowsAuto-check passed
Looper
ksimback/looper
Scaffold a well-designed agent loop with best-practice coaching and a cross-model review council.
710 GitHub stars~2.7k tokensUpdated 2 mo ago
AI & LLM EngineeringAuto-check: notes

Questions about Add Benchmark

What does Add Benchmark do?

When should I use Add Benchmark?

Add Benchmark fits situations like: tasks that involve LLM evaluation; tasks that involve Project management; tasks that involve Agent instruction files.

How do I install Add Benchmark in Claude Code?

Run `npx skills add ai-twinkle/Eval --skill add-benchmark -a claude-code`. Or copy the skill folder (.claude/skills/add-benchmark in ai-twinkle/Eval) into .claude/skills/add-benchmark in your project. Claude Code loads it when a task matches its description.

How do I install Add Benchmark in Codex?

Run `npx skills add ai-twinkle/Eval --skill add-benchmark -a codex`. Or copy the skill folder (.claude/skills/add-benchmark in ai-twinkle/Eval) into .agents/skills/add-benchmark in your project. Codex loads it when a task matches its description.

Can I use Add Benchmark in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add ai-twinkle/Eval --skill add-benchmark -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/add-benchmark, .gemini/skills/add-benchmark, .github/skills/add-benchmark and .opencode/skills/add-benchmark in your project.

What does Add Benchmark need to run?

Going by SKILL.md and its folder, Add Benchmark needs the command-line tools its instructions call (gh and python3). Our summary lists: Python 3.

Does Add Benchmark access the network?

SKILL.md contains no URLs. Its commands use gh, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Add Benchmark safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Add Benchmark use?

Add Benchmark is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Add Benchmark use?

About 1.8k tokens (SKILL.md is roughly 7.1k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Add Benchmark?

Skills that share tags, products or a category with Add Benchmark: Context Audit (undefined-ui/second-brain-os, 1k stars), Docs Sync (Kiln-AI/Kiln, 5.2k stars), Create Simple Prompt (pnp/copilot-prompts, 892 stars) and Write Skill (druxt/druxt.js, 114 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Add Benchmark?

ai-twinkle (a GitHub organization) maintains it in ai-twinkle/Eval, which has 117 GitHub stars. The repository holds 2 skills in this directory. The repository was last updated on September 15, 2026.

Source: ai-twinkle/Eval on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.

Install Add Benchmark

Facts

At a glance

What it does

When your agent uses it

Example prompts

Requirements

Workflow steps

What it can do on your machine

Tool permissions

Runs code

Network

Credentials

Context cost

Safety

SKILL.md

新增評測 Benchmark

動手前先確認

第 0 步：先開 GitHub，再寫程式

第 1 步：example dataset

第 2 步：實作 Extractor + Scorer

Evaluator 需要新的資料流嗎？

需要多指標（不只 accuracy）？

移植第三方程式碼

第 3 步：config 範本

第 4 步：分數對比（§6.3）

第 5 步：速度對比（§6.4）

第 6 步：文件

同時更新 datasets/example/README.md（§6.5.1）

第 7 步：測試（§6.6）

第 8 步：合入前檢查

push 前：強制 reviewer agent（§13）

版號（§11）

Files

Compare with similar skills

Similar skills

Context Audit

Docs Sync

Create Simple Prompt

Write Skill

Neat-Freak Knowledge Closeout

Looper

More from ai-twinkle/Eval

Run Eval

Related

Questions about Add Benchmark