Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Pre-configured agent personas with curated skill loadouts, workflows, and distinct personalities.
.claude/skills/alirezarezvani-persona-based-agents/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -45% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 95% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -37% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 109% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -45% | 0% |
Pre-configured agent personas with curated skill loadouts, workflows, and distinct personalities.
A persona is an agent definition that goes beyond "use these skills." Each persona includes:
bashcp agents/personas/startup-cto.md ~/.claude/agents/ # Then: "Activate startup-cto mode"
bash./scripts/convert.sh --tool cursor # Personas convert to .cursor/rules/*.mdc
bash./scripts/install.sh --tool <your-tool>
| Persona | Emoji | Domain | Best For | |---------|-------|--------|----------| | Startup CTO | 🏗️ | Engineering + Strategy | Technical co-founders, architecture decisions, team building | | Growth Marketer | 🚀 | Marketing + Growth | Bootstrapped founders, content-led growth, launches | | Solo Founder | 🦄 | Cross-domain | One-person startups, side projects, MVP building |
| | Task Agents (agents/) | Personas (agents/personas/) | |---|---|---| | Focus | Task execution | Role embodiment | | Scope | Single domain | Cross-domain curated set | | Voice | Neutral/professional | Personality-driven with backstory | | Workflows | Single-step | Multi-step with decision points | | Use case | "Do this task" | "Think like this person" |
Both coexist. Use task agents for focused work, personas for ongoing collaboration.
See TEMPLATE.md for the format specification. Key elements:
yaml--- name: Agent Name description: What this agent does and when to activate it. color: blue # Agent color theme emoji: 🎯 # Single emoji identifier vibe: One sentence personality capture. tools: Read, Write, Bash, Grep, Glob ---
Follow the section structure (Identity → Mission → Rules → Capabilities → Workflows → Communication → Metrics → Advanced → Learning) for consistency with existing personas.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 12,132 | 2,470 | -80% | 1 | 1 | 0% | 2,026 | 1,109 | -45% | 0 | 0 | — |
case-02 | fail→pass | 8,711 | 17,134 | +97% | 1 | 1 | 0% | 1,478 | 2,886 | +95% | 0 | 0 | — |
case-03 | fail→pass | 10,882 | 3,694 | -66% | 1 | 1 | 0% | 1,821 | 1,153 | -37% | 0 | 0 | — |
case-04 | fail→pass | 3,358 | 2,454 | -27% | 1 | 1 | 0% | 539 | 1,127 | +109% | 0 | 0 | — |
case-05 | pass→pass | 4,541 | 1,630 | -64% | 1 | 1 | 0% | 793 | 909 | +15% | 0 | 0 | — |
case-06 | pass→pass | 13,282 | 10,000 | -25% | 1 | 1 | 0% | 2,200 | 2,086 | -5% | 0 | 0 | — |
case-07 | pass→pass | 12,889 | 2,822 | -78% | 1 | 1 | 0% | 1,766 | 1,190 | -33% | 0 | 0 | — |
case-08 | fail→pass | 14,924 | 2,834 | -81% | 1 | 1 | 0% | 2,091 | 1,149 | -45% | 0 | 0 | — |
case-09 | pass→pass | 10,351 | 2,533 | -76% | 1 | 1 | 0% | 1,805 | 1,103 | -39% | 0 | 0 | — |
case-10 | fail→pass | 10,739 | 2,245 | -79% | 1 | 1 | 0% | 1,467 | 1,037 | -29% | 0 | 0 | — |
case-11 | pass→pass | 15,489 | 8,454 | -45% | 1 | 1 | 0% | 2,084 | 1,912 | -8% | 0 | 0 | — |
case-12 | fail→pass | 10,248 | 3,266 | -68% | 1 | 1 | 0% | 1,843 | 1,091 | -41% | 0 | 0 | — |
case-13 | fail→fail | 14,773 | 10,901 | -26% | 1 | 1 | 0% | 2,502 | 2,471 | -1% | 0 | 0 | — |
case-14 | fail→pass | 11,235 | 2,630 | -77% | 1 | 1 | 0% | 1,551 | 1,167 | -25% | 0 | 0 | — |
case-15 | fail→pass | 6,026 | 3,045 | -49% | 1 | 1 | 0% | 854 | 1,036 | +21% | 0 | 0 | — |
case-16 | fail→pass | 6,207 | 2,712 | -56% | 1 | 1 | 0% | 1,065 | 984 | -8% | 0 | 0 | — |
case-17 | fail→pass | 9,641 | 1,770 | -82% | 1 | 1 | 0% | 1,402 | 936 | -33% | 0 | 0 | — |
case-18 | fail→pass | 7,645 | 1,864 | -76% | 1 | 1 | 0% | 1,117 | 982 | -12% | 0 | 0 | — |
case-19 | pass→pass | 5,679 | 1,524 | -73% | 1 | 1 | 0% | 633 | 877 | +39% | 0 | 0 | — |
case-20 | pass→pass | 9,544 | 9,540 | -0% | 1 | 1 | 0% | 1,791 | 2,411 | +35% | 0 | 0 | — |
case-21 | pass→pass | 13,209 | 10,519 | -20% | 1 | 1 | 0% | 2,461 | 2,641 | +7% | 0 | 0 | — |
case-22 | pass→pass | 11,369 | 10,557 | -7% | 1 | 1 | 0% | 2,407 | 2,964 | +23% | 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. The headline lift of +55 percentage points is the difference between those two pass rates over the 22 comparable cases.
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.