Install
$ agentstack add skill-ryuketsukami-ste-plain-writing-ste-plain-writing ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
STE plain writing
Rewrite technical prose with rules adapted from ASD-STE100 Simplified Technical English. Preserve the facts, intent, useful detail, formatting, and established author voice.
Choose a mode
- Use strict mode for procedures, safety text, error messages, runbooks, and instructions.
- Use STE-flavored mode for READMEs, API docs, PR text, release notes, and other general technical prose.
- Use strict mode when the user does not specify a mode and the text tells a reader what to do.
- Use STE-flavored mode for all other requests.
Do not use this skill for code, identifiers, command syntax, creative writing, or marketing copy that depends on a distinct voice.
Apply the workflow
- Read the full source before editing.
- Identify facts, requirements, warnings, commands, links, and structural elements that must remain.
- Rewrite the text with the selected mode.
- Compare the rewrite with the source. Restore any lost meaning or necessary detail.
- Run the bundled linter when the text is available as a file.
- Fix meaningful violations. Treat the linter as a diagnostic tool, not a certification system.
- Return only the requested deliverable unless the user asks for an audit or explanation.
Use plain words
- Use one name for one thing.
- Use a short common word when it keeps the exact meaning.
- Prefer
starttocommence,usetoutilize,helptofacilitate, andshowtodemonstrate. - Give each word one meaning within the document.
- Use American spelling unless the source or user requires another variety.
- Remove empty intensifiers and promotional claims such as
seamless,robust,powerful,cutting-edge, andworld-class. - Keep necessary technical terms. Define an unfamiliar term at its first use.
Make actors and actions clear
- Use active voice when the actor is known.
- Use a verb for an action. Write
analyze the log, notperform an analysis of the log. - Prefer a simple tense to an
-ingmain verb. - Remove stacked auxiliaries and modal hedges.
- State who or what performs each action.
- Preserve passive voice when the actor is unknown, irrelevant, or intentionally withheld.
Control sentences and structure
- Put one instruction in each sentence.
- In strict mode, keep instructions at 20 words or fewer and descriptions at 25 words or fewer.
- In STE-flavored mode, prefer short sentences but vary length when it improves flow.
- Put a condition before the action it controls.
- Use a numbered vertical list for a sequence.
- Keep one topic in each paragraph and no more than six sentences.
- Replace semicolons with periods or another clear construction.
- Expand contractions in strict mode. Preserve them in STE-flavored mode when they match the author voice.
- Remove fake transitions, padded summaries, vague attributions, and repeated conclusions.
Preserve substance
- Do not invent evidence, examples, measurements, quotes, or sources.
- Do not remove caveats, limitations, warnings, or exceptions to make the text shorter.
- Do not flatten a known author voice into generic corporate prose.
- Do not add personality when neutral technical prose is correct.
- Do not claim that text is human-written or undetectable.
Run the linter
From the skill directory, run:
python3 scripts/ste_lint.py path/to/draft.md
The score reports heuristic violations per 100 words. Compare versions by score delta. Inspect each finding before editing because code samples and deliberate style choices can create false positives.
Check the final text
- Does each sentence add information or direct an action?
- Does every pronoun have a clear referent?
- Did the rewrite preserve all facts and constraints?
- Can a shorter common word keep the same meaning?
- Can the known actor replace passive voice?
- Does each list contain a real set or sequence?
- Did any AI-shaped filler, forced rule of three, vague authority, or empty closer remain?
- In strict mode, did any sentence exceed its length limit?
This skill improves the form of technical prose. It cannot make weak ideas true or unsupported claims credible.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: Ryuketsukami
- Source: Ryuketsukami/ste-plain-writing
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.