ime-lexicon-optimize skillA
ime-lexicon-optimize is agent-read markdown (skill) from taxueseek/ime-lexicon: 合并多份词表，自动改明显英文笔误和「词对、拼音错」，剔除 v2u / 拼音当词。 吃不准的进待确认，不写回输入法。.
Indexed from public GitHub and served as immutable, content-addressed versions. Install it pinned to an exact SHA-256 with the mdr CLI, and every file is verified against the hash recorded here before it reaches your agent. The deterministic audit below grades the latest version, and the same file always earns the same grade.
What the file says
# optimize — 修正个人词库 父技能：`ime-lexicon`。本文件只改表。 ## 何时用 - 豆包 + 微信两库合成一份 - 纠 `agnet` / `clade` / `skils` 这类明显错误 - 删指定垃圾词（`xx但是`） - 不要在这里搜新行业词（走 build / 原 industry-dict） ## 自动改 vs 待确认 自动改（表在脚本 `TYPO_MAP`）： - `agnet→agent` `clade→claude` `gork/grokgrok→grok` `flsah→flash` - `shft→shift` `skilsub→skill` `prma→proma` `boos→boss` - 词是英文、拼音只差 1 个字母，且不是高频缩写（`skills/skils` 改，`skill/skil` 留） 不自动改：`ocv`、产品名、短前缀。进 `待确认.tsv`。 ## 命令 ```bash python3 "$HOME/.agents/skills/ime-lexicon/scripts/ime_lexicon.py" optimize \ --from 工具/豆包输入法/个人词库.tsv \ --from 工具/微信输入法/个人词库.tsv \ --drop 'xx但是' --drop 'xx进入设置' \ --out 工具/词库整合/优化 ``` 产出：`词库.tsv` `更正记录.tsv` `待确认.tsv`。 `--no-auto` 只合并不改字。 `--profile personal`（默认）：只砍「的了 / 的吧」、≥16 字句子、v2u。习惯词（含频次=1 的「示例词」）留。`--drop-fragments` 才再剔个人碎片。 `--profile industry`：再砍过长（≥10 字）、频次=1 残片、「的了」类、叠词。行业表默认狠。 `--budget N`：裁到 N 条。重复 (词,拼音) 留个人库（source 含 doubao/wetype/personal）。个人已超 N 则个人按频次留前 N。 ## 代理步骤 1. 跑 optimize，报 in / out / fixed / review / profile。 2. 待确认里点名的才再改，不替用户猜缩写。 3. 要写回 → 父技能 import staging，问一句 apply。整表超过豆包剩余槽就加 `--budget`。
Read the whole file at its exact version.
How to install
mdr add taxueseek/ime-lexicon/ime-lexicon-optimize@git:20260825.e960d94mdr add taxueseek/ime-lexicon/ime-lexicon-optimize@sha256:6fc587b00f02867dPin to a label to follow the author's releases, or to a sha256 to freeze the exact bytes forever. Either way the resolved hash is written to mdr.lock, and mdr install reproduces it on any machine.
[](https://markdownregistry.com/a/art_r7jaugwdtbmocpk7)
1 badge views in 30 days
Versions
| version | committed | commit | size | audit | |
|---|---|---|---|---|---|
| git:20260825.e960d94 latest | 2026-08-25 | e960d94 | 1,980 B | A | view · diff |
| git:20260825.cca86d0 | 2026-08-25 | cca86d0 | 1,981 B | A | view · diff |
| git:20260822.a9f9084 | 2026-08-22 | a9f9084 | 1,988 B | A | view |
Audit of the latest version
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (1980 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
taxueseek/ime-lexicon · 5 stars · license MIT · pushed 2026-09-20 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_r7jaugwdtbmocpk7 GET https://markdownregistry.com/api/v1/resolve?ref=taxueseek/ime-lexicon/ime-lexicon-optimize GET https://markdownregistry.com/api/v1/blob/6fc587b00f02867d7b43f6e06a0da10d1e103f74124f01ec8d23fc1ccb47b7ba
Your agent does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.