ste-writing
Write or rewrite prose in ASD-STE100-style Simplified Technical English: short, literal sentences anyone can read without guessing.
Skill by amarjeet
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/amarjeet-ste-writing
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: writing
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/ste-writing
What's inside
- Skill name
ste-writing- Description
- Write or rewrite prose in ASD-STE100-style Simplified Technical English: short, literal sentences anyone can read without guessing. Accepts a target (where) and a mode (how). Triggers: ste, use ste, simplified technical english, asd-ste100, write in ste, rewrite in ste, plain technical english, make this unambiguous, ste review.
- License
- MIT
- Instructions
- 47 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
SKILL.md (first lines)
# STE Writing
## Overview
Apply ASD-STE100 rules to prose at a target and in a mode that the user names.
## Instructions
1. Read the invocation for **where** and **how**. Use the defaults if a value is missing.
| Where (target) | Scope |
|---|---|
| `reply` (default) | Only the current answer |
| `session` | Every answer until the user says `ste off` |
| `<path or glob>` | Prose in those files: docs, comments, commit or PR text, UI strings |
| pasted text | Only the text that the user gives |
| How (mode) | Action |
|---|---|
| `write` (default for `reply`/`session`) | Write new content in STE |
| `rewrite` (default for files and text) | Change existing prose to STE. Keep the meaning and the facts |
| `review` | Do not edit. List each violation as `location · rule · fix` |
Also obey free-text context, for example "for on-call runbooks" or "readers are non-native speakers". It sets the audience, the tone and the terms.
2. Apply the six rules to every sentence:
1. Max 20 words per sentence. Max 25 words for a procedure step.
2. One word, one meaning. Do not use one word for two ideas.
3. Same term every time. Pick one name for each thing. Do not use synonyms ("data" stays "data", never "information").
4. Active voice. Name who does the action.
5. One idea per sentence. One instruction per step. Put the condition first: "If X, do Y."
6. Simple, approved words. Use common words and simple verb tenses. Remove filler ("essentially", "basically", "considerably", "in order to").
3. Keep these unchanged: code, commands, identifiers, paths, URLs, quotations, product names. Technical nouns are allowed ("cache", "database"). Define each one at first use if the audience may not know it.
4. For files, edit only prose. Keep the structure, Markdown, and line format. Keep a terms list for the run. Use it to make the terms the same across all files.
5. After `rewrite` or `review`, add one stats line: `N sentences · max W words · T terms fixed`.
## Example
Input: `/ste-writing reply` "Explain what a cache is."
Output:
> A **cache** is a small, fast memory. It keeps a copy of **data** that you use frequently. When you need the **data**, the system reads the **cache** first. This is faster than reading the **database**. The **database** does less work.From amarjeet/agent-skills, MIT.
More from amarjeet/agent-skills
- 0.570score
story-loop
amarjeet
Story-driven implementation for a repository: a YAML story tracker cut from a design document and optional formal specs, a coordinator session that spawns an…
SkillAFreshness: AClaude Code +9 - 0.567score
firecrawl
amarjeet
Use the installed Firecrawl CLI to search, scrape, crawl, map, parse, monitor, or interact with web content and local documents.
SkillAFreshness: AClaude Code +9 - 0.526score
glyph-canvas
amarjeet
Render user requests as text drawings or diagrams using only unicode glyphs (box drawing, block elements, arrows, geometric shapes).
SkillBFreshness: BClaude Code +9 - 0.526score
skill-creator
amarjeet
End-to-end system for creating, updating, and validating agent skills (Cursor, Claude, Codex, Grok).
SkillBFreshness: BClaude Code +9 - 0.524score
dgx-spark-training-layout
amarjeet
Structure a fine-tuning experiment (LoRA, QLoRA, SFT) so each one lives in its own directory, heavy artifacts land in shared locations rather than beside the…
SkillBFreshness: BClaude Code +9 - 0.524score
dgx-spark-layout
amarjeet
Keep model weights and compiler caches out of individual experiment directories on a DGX Spark (or any single-node CUDA host) by pinning them to each tool's…
SkillBFreshness: BClaude Code +9 - 0.516score
dad-jokes
amarjeet
Tell a single dad joke on request.
SkillBFreshness: BClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Oct 6, 2026 (22 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/amarjeet/agent-skills
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from amarjeet
- 0.570score
story-loop
amarjeet
Story-driven implementation for a repository: a YAML story tracker cut from a design document and optional formal specs, a coordinator session that spawns an…
SkillAFreshness: AClaude Code +9 - 0.567score
firecrawl
amarjeet
Use the installed Firecrawl CLI to search, scrape, crawl, map, parse, monitor, or interact with web content and local documents.
SkillAFreshness: AClaude Code +9 - 0.526score
glyph-canvas
amarjeet
Render user requests as text drawings or diagrams using only unicode glyphs (box drawing, block elements, arrows, geometric shapes).
SkillBFreshness: BClaude Code +9 - 0.526score
skill-creator
amarjeet
End-to-end system for creating, updating, and validating agent skills (Cursor, Claude, Codex, Grok).
SkillBFreshness: BClaude Code +9 - 0.524score
dgx-spark-training-layout
amarjeet
Structure a fine-tuning experiment (LoRA, QLoRA, SFT) so each one lives in its own directory, heavy artifacts land in shared locations rather than beside the…
SkillBFreshness: BClaude Code +9 - 0.524score
dgx-spark-layout
amarjeet
Keep model weights and compiler caches out of individual experiment directories on a DGX Spark (or any single-node CUDA host) by pinning them to each tool's…
SkillBFreshness: BClaude Code +9
Cite this page
ste-writing skill. Agent Mart. https://agentmart.dev/skills/amarjeet-ste-writing (last updated 2026-10-24)