Skip to content
SkillModel-led1,910 tokens in use

dbs-speak-human

Rewrites Chinese messages meant for real people, such as WeChat chats, customer replies and agent answers, so they read like something a person would type.

11,518 stars on the repository Custom / Unknowndontbesilent2025/dbskill · skills/dbs-speak-human
npx -y skills add dontbesilent2025/dbskill -g --all
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing dbs-speak-human costs

204tokens idle
1,910tokens in use
4,895tokens, up to

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Up to adds the files its instructions link. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

Nothing detected: no shell, network, file writes, git or credentials in its files.

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Fair
  • Frontmatter8/15
  • Structure16/25
  • Examples9/15
  • Output0/15
  • Size13/15
  • Resources7/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent

No findings.

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Excellent
  • Development activity96
  • Maintenance100
  • Community64
  • Maturity28
  • Momentum40

dbs-speak-human writes or rewrites a Chinese message for a real reader so it sounds like casual speech. It removes written-style punctuation, list formatting, customer-service phrases and praise words that stand in for content, while keeping the facts, numbers and the speaker’s stance. If you supply messages you wrote yourself, it follows your tone.

It first works out who is speaking, to whom, where and what state the reader is in, then applies a checklist of common problems, self-checks each line by reading it as speech, and returns the full text with a short list of changes. After it has been called, the agent’s own replies in that chat follow the same standard.

What You Get

  • A rewritten message that can be copied straight into a chat
  • A change list pairing each original sentence with its new form and the reason
  • A note on which sentences were already natural and left untouched
  • A cleaner tone for the agent’s own replies for the rest of the conversation

Common Use Cases

  • Replying to a customer on WeChat without sounding like a script
  • Opening a group chat or sending a direct message that does not pressure the reader
  • Turning an agent’s formal answer into a short, direct reply

What Makes It Unique

It ships a detailed checklist built from real editing corrections: reader boundaries, stiff questions, written punctuation, customer-service words and a table of evaluation characters that hide missing detail. It replaces such words with the concrete fact or removes the clause, and it will not invent facts to fill the gap. It edits the text directly and leaves detection reports to other skills.

Under The Hood

The skill is one instruction file plus one checklist reference. It reads the draft and any samples of your own writing that you provide, and replies in the chat. It runs no scripts, makes no network calls and writes no files. The only side effect is that, once invoked, the agent applies the same speaking style to its later replies in that conversation.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers