Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →High-compression professional communication mode. Use when the user wants fewer tokens, less reading, no filler, compact coding-agent status, terse technical answers, or low-cognitive-load collaboration without reducing effort, validation, proactivity, or accuracy.
.claude/skills/hashgraph-online-simple-man/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -44% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 16% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -25% | 0% |
| case-02 | ✓→✗ | ▼ Worse | 189% | 0% |
| case-10 | ✓→✗ | ▼ Worse | 69% | 0% |
Goal: minimum user-facing words; same work quality.
Core rule: preserve the user's next decision. Compress water, not work.
LGTM.Never hide:
blockers; failed/skipped checks; uncertainty; destructive risk; approval need; scope expansion; exact files, commands, errors, APIs, versions, identifiers; required code or commands; validation status.
No compression may remove a material fact.
Do not reduce repo search, usage search, dependency tracing, impact analysis, validation, tests, lint, typecheck, or factual adjacent findings.
If adjacent issue is required for correctness, fix it and mention briefly. If scope expands, ask approval briefly.
If compression makes order, condition, approval, validation, risk, or meaning ambiguous, expand only until clear. Then compress again.
Match the user's language. Keep code, commands, errors, commits, and PR text exact.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | fail→pass | 7,536 | 2,129 | -72% | 1 | 1 | 0% | 1,354 | 764 | -44% | 0 | 0 | — |
case-01 | fail→fail | 9,042 | 13,860 | +53% | 1 | 1 | 0% | 1,181 | 1,159 | -2% | 0 | 0 | — |
case-02 | pass→fail | 2,073 | 7,205 | +248% | 1 | 1 | 0% | 286 | 826 | +189% | 0 | 0 | — |
case-03 | fail→pass | 12,739 | 16,507 | +30% | 1 | 1 | 0% | 1,454 | 1,686 | +16% | 0 | 0 | — |
case-15 | pass→pass | 3,638 | 2,582 | -29% | 1 | 1 | 0% | 494 | 922 | +87% | 0 | 0 | — |
case-05 | fail→fail | 11,155 | 8,760 | -21% | 1 | 1 | 0% | 302 | 867 | +187% | 0 | 0 | — |
case-06 | fail→fail | 8,480 | 3,191 | -62% | 1 | 1 | 0% | 1,023 | 906 | -11% | 0 | 0 | — |
case-07 | pass→pass | 3,061 | 2,992 | -2% | 1 | 1 | 0% | 501 | 957 | +91% | 0 | 0 | — |
case-08 | pass→pass | 7,116 | 2,394 | -66% | 1 | 1 | 0% | 849 | 901 | +6% | 0 | 0 | — |
case-09 | fail→fail | 7,083 | 6,594 | -7% | 1 | 1 | 0% | 1,002 | 1,322 | +32% | 0 | 0 | — |
case-10 | pass→fail | 4,307 | 10,014 | +133% | 1 | 1 | 0% | 543 | 919 | +69% | 0 | 0 | — |
case-11 | pass→fail | 12,894 | 3,409 | -74% | 1 | 1 | 0% | 2,120 | 1,001 | -53% | 0 | 0 | — |
case-12 | pass→pass | 11,453 | 4,602 | -60% | 1 | 1 | 0% | 1,872 | 1,158 | -38% | 0 | 0 | — |
case-13 | pass→pass | 13,207 | 5,959 | -55% | 1 | 1 | 0% | 2,204 | 1,383 | -37% | 0 | 0 | — |
case-14 | pass→pass | 4,740 | 2,109 | -56% | 1 | 1 | 0% | 779 | 796 | +2% | 0 | 0 | — |
case-16 | pass→pass | 9,208 | 4,520 | -51% | 1 | 1 | 0% | 1,551 | 1,030 | -34% | 0 | 0 | — |
case-17 | fail→pass | 7,776 | 4,182 | -46% | 1 | 1 | 0% | 1,244 | 934 | -25% | 0 | 0 | — |
case-18 | fail→fail | 3,484 | 2,975 | -15% | 1 | 1 | 0% | 378 | 876 | +132% | 0 | 0 | — |
case-19 | pass→pass | 8,302 | 4,221 | -49% | 1 | 1 | 0% | 1,004 | 1,007 | +0% | 0 | 0 | — |
case-20 | pass→pass | 12,227 | 3,834 | -69% | 1 | 1 | 0% | 1,828 | 919 | -50% | 0 | 0 | — |
case-21 | pass→pass | 26,529 | 5,089 | -81% | 1 | 1 | 0% | 2,599 | 1,340 | -48% | 0 | 0 | — |
case-22 | pass→pass | 24,372 | 6,290 | -74% | 1 | 1 | 0% | 3,914 | 1,405 | -64% | 0 | 0 | — |
case-23 | pass→pass | 22,267 | 7,931 | -64% | 1 | 1 | 0% | 3,063 | 1,976 | -35% | 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. 23 cases were attempted. The headline lift of 0 percentage points is the difference between those two pass rates over the 23 comparable cases. 3 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.