Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Browse, search, and contribute community hacks — real-world patterns for provider switching, teams, skills, hooks, cost optimization, and more.
.claude/skills/automagik-dev-genie-hacks/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | -4% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 264% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 2% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 32% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 21% | 0% |
Runtime syntax: invoke the plugin copy through the active runtime's owner-qualified skill selector; use a bare selector only when intentionally selecting a user-tier copy (a separately installed personal copy; Genie no longer seeds this tier). Cross-skill prose below uses bare names as portable semantic routes; the orchestrator resolves the selector for the active runtime.
Browse real-world Genie patterns contributed by the community. Search by problem, explore by category, or contribute your own.
genie-hacks with any subcommand (no subcommand → list)references/catalog.md (relative to this skill's directory) before answering any list/search/show/help request — it holds every hack (problem, solution, code, benefit, when-to-use) plus the category table. Never invent a hack that isn't in it.references/contributing.md when running contribute — exact fork/branch/PR commands and error recovery.genie/hacks.mdx in automagik-dev/docs).| Command | Behavior | |---------|----------| | genie-hacks / list | Table of all hacks — ID, title, category — then the count and a contribute nudge. | | search <keyword> | Case-insensitive match over title/problem/solution/code; show ID, title, category, problem snippet per match. None → suggest broader terms or list. | | show <hack-id> | Full entry: problem, solution, code, benefit, when to use. Unknown ID → suggest the closest IDs. | | help <problem> | Match the described problem to the top 3 relevant hacks; one-line "why" plus a quick tip each. Prefer a loose match over "no matches". | | contribute | Guided submission → automated PR to automagik-dev/docs. |
End list and show with a pointer to show <id> and contribute.
references/contributing.md: preflight GitHub access, fork/clone or use the GitHub connector, branch hack/<slug>, append under the category heading in genie/hacks.mdx, commit hack: <title>, and open the PR against dev.If gh is missing/unauthenticated or any GitHub step fails, save the formatted hack to ~/.genie/cache/pending-hacks/<hack-id>.md and relay the manual PR steps — never lose the user's write-up.
genie --help is the source of truth); v4-era entries carry a note with the live replacement.dev — never main/master.list, full format only for show.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 15,126 | 4,701 | -69% | 1 | 1 | 0% | 2,213 | 1,150 | -48% | 0 | 0 | — |
case-02 | fail→fail | 14,117 | 6,420 | -55% | 1 | 1 | 0% | 2,318 | 1,168 | -50% | 0 | 0 | — |
case-03 | fail→fail | 13,989 | 5,626 | -60% | 1 | 1 | 0% | 2,145 | 1,813 | -15% | 0 | 0 | — |
case-04 | fail→fail | 12,812 | 2,596 | -80% | 1 | 1 | 0% | 1,973 | 1,157 | -41% | 0 | 0 | — |
case-05 | fail→pass | 16,247 | 13,173 | -19% | 1 | 1 | 0% | 2,575 | 2,482 | -4% | 0 | 0 | — |
case-06 | fail→pass | 3,416 | 7,058 | +107% | 1 | 1 | 0% | 438 | 1,595 | +264% | 0 | 0 | — |
case-07 | fail→fail | 2,968 | 3,112 | +5% | 1 | 1 | 0% | 343 | 1,315 | +283% | 0 | 0 | — |
case-08 | fail→pass | 9,571 | 3,876 | -60% | 1 | 1 | 0% | 1,446 | 1,468 | +2% | 0 | 0 | — |
case-09 | fail→pass | 8,211 | 5,517 | -33% | 1 | 1 | 0% | 1,315 | 1,740 | +32% | 0 | 0 | — |
case-10 | fail→fail | 6,513 | 4,290 | -34% | 1 | 1 | 0% | 1,011 | 1,578 | +56% | 0 | 0 | — |
case-11 | fail→fail | 6,508 | 2,653 | -59% | 1 | 1 | 0% | 1,011 | 1,306 | +29% | 0 | 0 | — |
case-12 | fail→pass | 6,610 | 1,590 | -76% | 1 | 1 | 0% | 932 | 1,127 | +21% | 0 | 0 | — |
case-13 | pass→pass | 6,642 | 3,040 | -54% | 1 | 1 | 0% | 1,062 | 1,405 | +32% | 0 | 0 | — |
case-14 | fail→pass | 10,826 | 4,672 | -57% | 1 | 1 | 0% | 1,612 | 1,576 | -2% | 0 | 0 | — |
case-15 | pass→pass | 9,923 | 2,664 | -73% | 1 | 1 | 0% | 1,557 | 1,231 | -21% | 0 | 0 | — |
case-16 | pass→pass | 9,527 | 2,428 | -75% | 1 | 1 | 0% | 1,561 | 1,273 | -18% | 0 | 0 | — |
case-17 | fail→pass | 7,688 | 1,763 | -77% | 1 | 1 | 0% | 1,191 | 1,117 | -6% | 0 | 0 | — |
case-18 | fail→pass | 8,705 | 2,095 | -76% | 1 | 1 | 0% | 1,328 | 1,185 | -11% | 0 | 0 | — |
case-19 | pass→pass | 11,191 | 4,695 | -58% | 1 | 1 | 0% | 1,607 | 1,616 | +1% | 0 | 0 | — |
case-20 | pass→fail | 7,961 | 10,400 | +31% | 1 | 1 | 0% | 1,341 | 2,158 | +61% | 0 | 0 | — |
case-21 | pass→fail | 16,442 | 12,998 | -21% | 1 | 1 | 0% | 2,968 | 2,548 | -14% | 0 | 0 | — |
case-22 | pass→pass | 6,512 | 6,344 | -3% | 1 | 1 | 0% | 1,146 | 1,493 | +30% | 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. 22 cases were attempted, and 20 counted toward the lift figure. The other 2 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 +27 percentage points is the difference between those two pass rates over the 20 comparable cases. 2 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.