Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when writing Go code with `github.com/ax-llm/ax/packages/go` for agent memory, recall callbacks, dynamic skill discovery, loaded-skill state, and used-skill tracking.
.claude/skills/ax-llm-ax-go-agent-memory-skills/SKILL.md| Model | Eval pass | Runs |
|---|---|---|
| gemini-3.6-flash | 100% | 20 |
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -24% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 10% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -20% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -49% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -64% | 0% |
This skill helps an agent write Go code with the generated Ax package github.com/ax-llm/ax/packages/go. Use the generated package API, examples, and manifests; do not import TypeScript-only APIs unless you are editing the TypeScript package.
github.com/ax-llm/ax/packages/go.API.md and axir-api.json.axir-capabilities.json.examples/.javascript-goja.gohelper := ax.NewAgent("question:string -> answer:string", nil) out := helper.Forward(llm, map[string]ax.Value{"question": "How should I proceed?"}, nil)
skills seed the loaded-skill prompt without firing load observers.skills override constructor entries by normalized ID and remain loaded for later calls. IDs and names are trimmed, malformed entries are skipped, valid empty content is preserved, and rendered entries are ID-sorted.memories seeds the first actor turn. Recalled entries merge by ID for that run, then memory state resets before the next forward.onSkillsSearch / onMemoriesSearch take precedence over static catalogs. Without a host callback, skillsCatalog / memoriesCatalog use the built-in deterministic lexical ranker.onLoadedMemories / onLoadedSkills observe runtime recall and discovery, not constructor presets. onUsedMemories / onUsedSkills emit one consolidated notification per forward. Forward observers override constructor observers, and observer errors are ignored.relevanceRanking produces advisory skill and memory hints using the same tokenization, weighting, tie suppression, limits, snippets, and already-loaded exclusion as TypeScript.GetState() and SetState(...) preserve the legacy bare-runtime snapshot shape. Use ExportRuntimeState() and RestoreRuntimeState(...) for the complete portable agent snapshot, including loaded skills and constructor-preset reapplication. Do not interchange the two shapes.src/examples/go/long-agents/skills_and_memory_assistant.go.smart-defaults-agent example under src/examples/go/long-agents/.axllm.NewAgent, axllm.AxAgent, AxAgent.AddChildAgentaxllm.ProcessCodeRuntime, axllm.RuntimeCapabilities, axllm.RuntimeEnvelope, javascript-gojaprovider-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 | 26,001 | 17,532 | -33% | 1 | 1 | 0% | 4,176 | 3,174 | -24% | 0 | 0 | — |
case-02 | fail→pass | 21,936 | 18,626 | -15% | 1 | 1 | 0% | 3,349 | 3,672 | +10% | 0 | 0 | — |
case-03 | fail→pass | 32,051 | 22,440 | -30% | 1 | 1 | 0% | 5,107 | 4,072 | -20% | 0 | 0 | — |
case-04 | pass→pass | 18,495 | 15,047 | -19% | 1 | 1 | 0% | 2,226 | 2,516 | +13% | 0 | 0 | — |
case-05 | pass→pass | 19,636 | 13,642 | -31% | 1 | 1 | 0% | 2,524 | 2,332 | -8% | 0 | 0 | — |
case-06 | pass→pass | 15,650 | 13,546 | -13% | 1 | 1 | 0% | 1,831 | 2,503 | +37% | 0 | 0 | — |
case-07 | fail→pass | 25,972 | 6,883 | -73% | 1 | 1 | 0% | 2,350 | 1,204 | -49% | 0 | 0 | — |
case-08 | fail→pass | 22,936 | 6,526 | -72% | 1 | 1 | 0% | 3,093 | 1,118 | -64% | 0 | 0 | — |
case-09 | fail→pass | 24,445 | 6,301 | -74% | 1 | 1 | 0% | 3,182 | 1,073 | -66% | 0 | 0 | — |
case-10 | fail→pass | 22,722 | 6,850 | -70% | 1 | 1 | 0% | 2,739 | 1,096 | -60% | 0 | 0 | — |
case-11 | fail→pass | 18,096 | 8,178 | -55% | 1 | 1 | 0% | 2,024 | 1,368 | -32% | 0 | 0 | — |
case-12 | pass→pass | 16,308 | 7,660 | -53% | 1 | 1 | 0% | 1,733 | 1,283 | -26% | 0 | 0 | — |
case-13 | fail→pass | 19,780 | 9,237 | -53% | 1 | 1 | 0% | 2,459 | 1,501 | -39% | 0 | 0 | — |
case-14 | fail→pass | 18,071 | 8,597 | -52% | 1 | 1 | 0% | 2,010 | 1,520 | -24% | 0 | 0 | — |
case-15 | pass→pass | 16,695 | 7,337 | -56% | 1 | 1 | 0% | 1,925 | 1,328 | -31% | 0 | 0 | — |
case-16 | fail→pass | 13,618 | 11,730 | -14% | 1 | 1 | 0% | 1,485 | 1,995 | +34% | 0 | 0 | — |
case-17 | pass→pass | 19,598 | 7,884 | -60% | 1 | 1 | 0% | 2,233 | 1,359 | -39% | 0 | 0 | — |
case-18 | fail→pass | 14,353 | 7,441 | -48% | 1 | 1 | 0% | 1,522 | 1,227 | -19% | 0 | 0 | — |
case-19 | fail→pass | 20,956 | 18,376 | -12% | 1 | 1 | 0% | 2,674 | 2,985 | +12% | 0 | 0 | — |
case-20 | fail→pass | 25,583 | 7,957 | -69% | 1 | 1 | 0% | 4,115 | 1,365 | -67% | 0 | 0 | — |
case-21 | pass→pass | 16,292 | 7,973 | -51% | 1 | 1 | 0% | 1,892 | 1,383 | -27% | 0 | 0 | — |
case-22 | fail→pass | 14,460 | 7,569 | -48% | 1 | 1 | 0% | 1,548 | 1,231 | -20% | 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 +68 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 | 8/11/2026 | +36% |
Other measured skills in the registry, with their headline benchmark lift.