Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when writing Rust code with `axllm` for agent tracing, centralized and multi-tenant usage accounting, action logs, runtime diagnostics, replay, and production debugging.
.claude/skills/ax-llm-ax-rust-agent-observability/SKILL.md| Model | Eval pass | Runs |
|---|---|---|
| gemini-3.6-flash | 100% | 30 |
| gemini-3.1-pro-preview | 100% | 1 |
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -16% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 25% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 2% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 0% | 0% |
This skill helps an agent write Rust code with the generated Ax package axllm. Use the generated package API, examples, and manifests; do not import TypeScript-only APIs unless you are editing the TypeScript package.
axllm.API.md and axir-api.json.axir-capabilities.json.examples/.javascript-quickjs.rustlet helper = axllm::agent("question:string -> answer:string")?; let out = helper.forward(&llm, inputs, None)?;
Use the process-wide usage observer for application accounting across many agents, API routes, tenants, and users. Keep per-agent usage accessors for inspecting one agent instance after a run.
rustset_usage_observer(Some(Arc::new(move |event| { usage_queue.push(event); }))); // Later: set_usage_observer(None);
usageContext in AI service options for stable application or environment defaults. Attach it in call or agent-forward option maps for tenant, user, request, run, and feature attribution.attributes are shallow-merged.src/examples/rust/generation/usage_observer.rs.ai, typesafe, AxAITypesafeClient, AxCancellationToken, AxAIServiceAbortedError, get_supported_ai_models, AxCredentialRequest, AxCredentialProvider, AxAIClient::owned_worker_factory, AxChatSession, AxChatStream, OpenAICompatibleClient, OpenAIResponsesClient, GoogleGeminiClient, AnthropicClient, AxUsageContext, AxUsageEvent, AxUsageObserver, set_usage_observer, AxRuntimeHooks, AxRateLimitInfo, AxRateLimiter, AxTracer, AxMeter, AxGlobals, set_rate_limiter, set_tracer, set_meter, AxBalancer, AxBalancerAdaptiveStrategy, AxBalancerStatsStore, AxInMemoryBalancerStatsStore, create_balancer_route_stats, update_balancer_route_stats, sample_balancer_route_health, MultiServiceRouter, ProviderRouteragent, AxAgent, AxAgent::with_child_agentProcessCodeRuntime, RuntimeCapabilities, RuntimeEnvelope, javascript-quickjsprovider-api examples only when the user explicitly has provider credentials available.no-key examples for deterministic local checks and provider request mapping.tools/*/skills/ into user packages.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 43,877 | 20,147 | -54% | 1 | 1 | 0% | 3,563 | 3,863 | +8% | 0 | 0 | — |
case-02 | fail→pass | 22,584 | 18,951 | -16% | 1 | 1 | 0% | 3,833 | 3,222 | -16% | 0 | 0 | — |
case-03 | fail→pass | 26,268 | 12,681 | -52% | 1 | 1 | 0% | 2,506 | 3,127 | +25% | 0 | 0 | — |
case-04 | fail→pass | 20,572 | 13,229 | -36% | 1 | 1 | 0% | 3,148 | 3,220 | +2% | 0 | 0 | — |
case-05 | fail→pass | 23,025 | 10,872 | -53% | 1 | 1 | 0% | 2,911 | 2,904 | -0% | 0 | 0 | — |
case-06 | fail→pass | 17,486 | 11,213 | -36% | 1 | 1 | 0% | 2,301 | 2,940 | +28% | 0 | 0 | — |
case-07 | fail→pass | 11,438 | 7,802 | -32% | 1 | 1 | 0% | 1,258 | 2,360 | +88% | 0 | 0 | — |
case-08 | fail→pass | 13,774 | 10,672 | -23% | 1 | 1 | 0% | 1,924 | 1,791 | -7% | 0 | 0 | — |
case-09 | pass→pass | 17,758 | 7,838 | -56% | 1 | 1 | 0% | 2,353 | 2,176 | -8% | 0 | 0 | — |
case-10 | pass→pass | 11,548 | 7,242 | -37% | 1 | 1 | 0% | 1,634 | 1,812 | +11% | 0 | 0 | — |
case-11 | fail→pass | 11,260 | 7,219 | -36% | 1 | 1 | 0% | 1,603 | 1,728 | +8% | 0 | 0 | — |
case-12 | fail→pass | 14,672 | 10,967 | -25% | 1 | 1 | 0% | 2,314 | 2,694 | +16% | 0 | 0 | — |
case-13 | pass→pass | 15,554 | 6,819 | -56% | 1 | 1 | 0% | 2,359 | 1,888 | -20% | 0 | 0 | — |
case-14 | pass→pass | 16,173 | 11,173 | -31% | 1 | 1 | 0% | 2,292 | 3,095 | +35% | 0 | 0 | — |
case-15 | pass→pass | 10,934 | 5,252 | -52% | 1 | 1 | 0% | 1,348 | 1,596 | +18% | 0 | 0 | — |
case-16 | fail→pass | 20,275 | 8,080 | -60% | 1 | 1 | 0% | 1,539 | 1,926 | +25% | 0 | 0 | — |
case-17 | fail→pass | 28,462 | 18,734 | -34% | 1 | 1 | 0% | 2,561 | 3,008 | +17% | 0 | 0 | — |
case-18 | fail→pass | 14,922 | 9,861 | -34% | 1 | 1 | 0% | 1,598 | 2,462 | +54% | 0 | 0 | — |
case-19 | fail→pass | 22,328 | 8,486 | -62% | 1 | 1 | 0% | 2,084 | 2,058 | -1% | 0 | 0 | — |
case-20 | pass→pass | 17,337 | 7,709 | -56% | 1 | 1 | 0% | 2,487 | 2,153 | -13% | 0 | 0 | — |
case-21 | pass→pass | 12,422 | 4,055 | -67% | 1 | 1 | 0% | 1,292 | 1,452 | +12% | 0 | 0 | — |
case-22 | fail→fail | 14,023 | 4,456 | -68% | 1 | 1 | 0% | 1,753 | 1,631 | -7% | 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 +64 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 9/9/2026 | +68% |
| gemini-3.6-flash | verified | 9/2/2026 | +45% |
| gemini-3.6-flash | verified | 8/26/2026 | +41% |
| gemini-3.6-flash | verified | 8/17/2026 | +64% |
| gemini-3.6-flash | verified | 8/10/2026 | +45% |
Other measured skills in the registry, with their headline benchmark lift.