Apply clear-writing standards to any prose the agent produces - READMEs, docs, UI copy, error messages, commit and PR text, release notes.
SkillScore86Maintained
Automatic
◆Writing & Editing pack
Last verified 8/13/2026
What it changes
Passed 19 of 22 test cases with the skill, and 18 of 22 without it.
Without skill82%
With skill86%
Checks it passed that the model alone missed
The rewritten response uses prose sentences instead of a numbered list
The output provides a single rewritten version rather than multiple options
The response leads directly with the tightened text without conversational preamble
2 fixed17 passed either way2 still failing1 broke
Pass-rate lift
+5 pp
Tokens
+26%
Agent turns
0%
Response time
-28%
Model gemini-3.6-flashMeasured 8/13/2026v119 of 22 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 2 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 2
Prompt
Edit this list from an email update: 'Steps completed: 1. Upgraded Node. 2. Restarted cluster.' Rewrite this as a single clean prose sentence.
What was checked, without and with the skill
Checked
Without skill
With skill
The rewritten response uses prose sentences instead of a numbered list
Missed
Passed
case-19 · gemini-3.6-flash · 8/13/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md43 linessingle file
writing-guidelines
Write for a reader who is busy and did not see your work. Readability beats
cleverness; clarity beats completeness.
Rules
Lead with the outcome. The first sentence answers what happened or what
the thing is - the line the reader would ask for if they said "just the TLDR".
Supporting detail comes after.
Cut filler. Delete "just", "simply", "basically", "in order to", "it is
important to note". If a sentence changes nothing when removed, remove it.
Concrete over abstract. Name the file, the number, the command. "Faster"
is weaker than "cuts the build from 40s to 9s".
Active voice, one idea per sentence. Short sentences that each carry one
point read faster than long ones that carry three.
Consistent terms. Use the project's own vocabulary, the same word for the
same thing every time. Tie names to the shared-language CONTEXT.md when one
exists (see domain-modeling).
Readable, not clipped. Being short and being clear are different. Achieve
short by dropping what the reader does not need, not by compressing prose into
fragments, arrow chains (A -> B -> fails), or invented abbreviations.
Error messages and UI copy
Say what happened and what to do next: Config not found at ./app.config.ts. Create it or pass --config <path>. beats Error: config missing.
No dead ends. Every failure names a next step.
Match the product's voice; drop exclamation marks and filler enthusiasm.
AI-slop tells to strip
Em-dashes and en-dashes - use a spaced hyphen or a full stop. This is the most
reliable machine-written tell.
Hollow openers: "In today's fast-paced world", "It's worth noting that".
Hedging stacks: "might potentially perhaps".
Binary flourishes: "Not X. But Y." as a rhetorical beat.
Applies the scientific method to debugging by helping users form specific, testable hypotheses, design targeted experiments, and systematically confirm or reject theories to find root causes. Use when a user says their code isn't working, they're getting an error, something broke, they want to troubleshoot a bug, or they're trying to figure out what's causing an issue. Concrete actions include isolating failing components, forming and testing hypotheses, analyzing error messages, tracing executi