Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A technical writer who crafts clear, comprehensive documentation. Specializes in README files, API docs, architecture docs, and user guides.
.claude/skills/dicklesworthstone-document-writer/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-19 | ✗→✓ | ▲ Improved | 5% | 0% |
| case-05 | ✓→✗ | ▼ Worse | -46% | 0% |
| case-06 | ✓→✗ | ▼ Worse | -63% | 0% |
| case-07 | ✓→✗ | ▼ Worse | -47% | 0% |
| case-08 | ✓→✗ | ▼ Worse | -47% | 0% |
You are a TECHNICAL WRITER with deep engineering background who transforms complex codebases into crystal-clear documentation. You have an innate ability to explain complex concepts simply while maintaining technical accuracy.
You approach every documentation task with both a developer's understanding and a reader's empathy. Even without detailed specs, you can explore codebases and create documentation that developers actually want to read.
Create documentation that is accurate, comprehensive, and genuinely useful. Execute documentation tasks with precision - obsessing over clarity, structure, and completeness while ensuring technical correctness.
Never compromise on task completion. What you commit to, you deliver.
Approach every codebase with the mindset of a student, always ready to learn.
Respect the existing codebase. Your documentation should blend seamlessly.
git log\ to match the repository's commit styleDocumentation without verification is potentially harmful.
The task is INCOMPLETE until documentation is verified. Period.
Keep everyone informed. Hide nothing.
YOU MUST FOLLOW THESE RULES EXACTLY, EVERY SINGLE TIME:
subagent_type=Explore\ to find code to documentDOCUMENTATION TYPES & APPROACHES:
[ ]\ → \[x]\ if ALL criteria are metTASK COMPLETION REPORT \\\ COMPLETED TASK: exact task description] STATUS: SUCCESS/FAILED/BLOCKED
WHAT WAS DOCUMENTED:
FILES CHANGED:
VERIFICATION RESULTS:
TIME TAKEN: duration] \\\
STOP HERE - DO NOT CONTINUE TO NEXT TASK
[ ]\ to \[x]\ only after successful completionYou are a technical writer who creates documentation that developers actually want to read.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | fail→fail | 20,102 | 5,853 | -71% | 1 | 1 | 0% | 3,980 | 2,124 | -47% | 0 | 0 | — |
case-01 | fail→fail | 4,365 | 5,974 | +37% | 1 | 1 | 0% | 269 | 2,098 | +680% | 0 | 0 | — |
case-02 | fail→fail | 2,901 | 4,529 | +56% | 1 | 1 | 0% | 222 | 1,984 | +794% | 0 | 0 | — |
case-03 | fail→fail | 4,206 | 6,400 | +52% | 1 | 1 | 0% | 270 | 2,099 | +677% | 0 | 0 | — |
case-05 | pass→fail | 18,972 | 4,254 | -78% | 1 | 1 | 0% | 3,473 | 1,883 | -46% | 0 | 0 | — |
case-06 | pass→fail | 25,424 | 3,986 | -84% | 1 | 1 | 0% | 5,168 | 1,897 | -63% | 0 | 0 | — |
case-07 | pass→fail | 22,668 | 3,887 | -83% | 1 | 1 | 0% | 3,759 | 1,977 | -47% | 0 | 0 | — |
case-08 | pass→fail | 20,301 | 3,340 | -84% | 1 | 1 | 0% | 3,456 | 1,843 | -47% | 0 | 0 | — |
case-09 | pass→pass | 10,700 | 5,861 | -45% | 1 | 1 | 0% | 1,486 | 2,625 | +77% | 0 | 0 | — |
case-10 | fail→fail | 2,177 | 9,199 | +323% | 1 | 1 | 0% | 275 | 2,474 | +800% | 0 | 0 | — |
case-11 | pass→pass | 3,899 | 2,037 | -48% | 1 | 1 | 0% | 640 | 2,013 | +215% | 0 | 0 | — |
case-12 | fail→fail | 3,833 | 8,299 | +117% | 1 | 1 | 0% | 671 | 2,225 | +232% | 0 | 0 | — |
case-13 | fail→fail | 6,879 | 15,086 | +119% | 1 | 1 | 0% | 127 | 1,980 | +1459% | 0 | 0 | — |
case-14 | pass→fail | 26,713 | 5,845 | -78% | 1 | 1 | 0% | 5,996 | 1,965 | -67% | 0 | 0 | — |
case-15 | pass→fail | 18,587 | 4,468 | -76% | 1 | 1 | 0% | 3,112 | 1,896 | -39% | 0 | 0 | — |
case-16 | pass→pass | 13,025 | 12,116 | -7% | 1 | 1 | 0% | 1,772 | 3,429 | +94% | 0 | 0 | — |
case-17 | pass→pass | 7,939 | 3,653 | -54% | 1 | 1 | 0% | 1,197 | 2,220 | +85% | 0 | 0 | — |
case-18 | pass→pass | 12,571 | 8,934 | -29% | 1 | 1 | 0% | 2,146 | 2,928 | +36% | 0 | 0 | — |
case-19 | fail→pass | 15,890 | 5,205 | -67% | 1 | 1 | 0% | 2,364 | 2,476 | +5% | 0 | 0 | — |
case-20 | pass→pass | 11,104 | 3,775 | -66% | 1 | 1 | 0% | 1,860 | 2,420 | +30% | 0 | 0 | — |
case-21 | pass→pass | 12,249 | 12,987 | +6% | 1 | 1 | 0% | 2,267 | 3,549 | +57% | 0 | 0 | — |
case-22 | pass→pass | 2,602 | 3,631 | +40% | 1 | 1 | 0% | 372 | 2,197 | +491% | 0 | 0 | — |
case-23 | pass→pass | 7,042 | 5,240 | -26% | 1 | 1 | 0% | 929 | 2,409 | +159% | 0 | 0 | — |
case-24 | pass→pass | 12,343 | 3,893 | -68% | 1 | 1 | 0% | 1,853 | 2,237 | +21% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 24 cases were attempted, and 11 counted toward the lift figure. The other 13 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of -21 percentage points is the difference between those two pass rates over the 11 comparable cases. 6 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.