Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create and adapt Dynamic Agentic Architecture agents that learn and evolve
.claude/skills/ruvnet-daa-agent/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-12 | ✗→✓ | ▲ Improved | -17% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -28% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -51% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -8% | 0% |
Create agents with Dynamic Agentic Architecture that adapt and learn over time.
When you need agents that go beyond static configurations — agents that adapt their behavior based on performance metrics, learn from interactions, and share knowledge with other agents.
mcp__plugin_ruflo-core_ruflo__daa_agent_create with initial configuration and learning parametersmcp__plugin_ruflo-core_ruflo__daa_learning_status to see adaptation progressmcp__plugin_ruflo-core_ruflo__daa_performance_metrics for efficiency and accuracy metricsmcp__plugin_ruflo-core_ruflo__daa_agent_adapt to trigger manual adaptation based on feedbackmcp__plugin_ruflo-core_ruflo__daa_knowledge_share to propagate learnings to other agents| Aspect | Static Agent | DAA Agent | |--------|-------------|-----------| | Behavior | Fixed configuration | Adapts over time | | Learning | None | Continuous from interactions | | Knowledge | Isolated | Shared across agents | | Performance | Constant | Improves with use |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-06 | pass→pass | 13,247 | 5,180 | -61% | 1 | 1 | 0% | 2,191 | 1,176 | -46% | 0 | 0 | — |
case-12 | fail→pass | 13,043 | 8,068 | -38% | 1 | 1 | 0% | 2,259 | 1,869 | -17% | 0 | 0 | — |
case-01 | fail→fail | 13,906 | 6,139 | -56% | 1 | 1 | 0% | 2,426 | 762 | -69% | 0 | 0 | — |
case-02 | fail→fail | 8,646 | 7,626 | -12% | 1 | 1 | 0% | 1,583 | 1,177 | -26% | 0 | 0 | — |
case-03 | fail→fail | 17,555 | 10,619 | -40% | 1 | 1 | 0% | 3,931 | 918 | -77% | 0 | 0 | — |
case-04 | pass→pass | 13,939 | 5,411 | -61% | 1 | 1 | 0% | 2,139 | 1,226 | -43% | 0 | 0 | — |
case-05 | pass→pass | 14,894 | 10,501 | -29% | 1 | 1 | 0% | 2,339 | 1,926 | -18% | 0 | 0 | — |
case-07 | pass→fail | 16,312 | 4,949 | -70% | 1 | 1 | 0% | 2,777 | 723 | -74% | 0 | 0 | — |
case-08 | fail→pass | 15,217 | 6,935 | -54% | 1 | 1 | 0% | 2,560 | 1,839 | -28% | 0 | 0 | — |
case-09 | fail→pass | 11,147 | 2,729 | -76% | 1 | 1 | 0% | 1,764 | 871 | -51% | 0 | 0 | — |
case-10 | fail→pass | 12,556 | 7,226 | -42% | 1 | 1 | 0% | 2,222 | 1,475 | -34% | 0 | 0 | — |
case-11 | fail→fail | 12,573 | 8,608 | -32% | 1 | 1 | 0% | 1,980 | 2,024 | +2% | 0 | 0 | — |
case-13 | fail→pass | 13,495 | 8,914 | -34% | 1 | 1 | 0% | 2,229 | 2,044 | -8% | 0 | 0 | — |
case-14 | fail→pass | 16,160 | 6,240 | -61% | 1 | 1 | 0% | 2,557 | 1,564 | -39% | 0 | 0 | — |
case-15 | fail→pass | 7,395 | 4,659 | -37% | 1 | 1 | 0% | 1,344 | 1,307 | -3% | 0 | 0 | — |
case-16 | fail→pass | 9,631 | 5,130 | -47% | 1 | 1 | 0% | 1,623 | 1,239 | -24% | 0 | 0 | — |
case-17 | pass→pass | 11,042 | 4,607 | -58% | 1 | 1 | 0% | 1,848 | 1,082 | -41% | 0 | 0 | — |
case-18 | fail→pass | 10,512 | 2,968 | -72% | 1 | 1 | 0% | 1,936 | 879 | -55% | 0 | 0 | — |
case-19 | pass→pass | 13,680 | 12,832 | -6% | 1 | 1 | 0% | 2,792 | 2,923 | +5% | 0 | 0 | — |
case-20 | pass→pass | 14,936 | 11,441 | -23% | 1 | 1 | 0% | 2,976 | 2,495 | -16% | 0 | 0 | — |
case-21 | pass→pass | 15,077 | 14,723 | -2% | 1 | 1 | 0% | 3,167 | 3,200 | +1% | 0 | 0 | — |
case-22 | pass→pass | 11,845 | 7,568 | -36% | 1 | 1 | 0% | 2,276 | 1,570 | -31% | 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 18 counted toward the lift figure. The other 4 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 +36 percentage points is the difference between those two pass rates over the 18 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.