{"record":{"id":"e07973bc695a43fe","repo":"multica-ai/multica","slug":"load-skill-bundle-cache-w","errorCode":null,"errorMessage":"load skill bundle cache: %w","messagePattern":"load skill bundle cache: %w","errorType":"exception","errorClass":null,"httpStatus":null,"severity":"error","filePath":"server/internal/daemon/daemon.go","lineNumber":5871,"sourceCode":"\nfunc (d *Daemon) ensureTaskSkillBundles(ctx context.Context, task *Task) error {\n\tif task == nil || task.Agent == nil || len(task.Agent.SkillRefs) == 0 {\n\t\treturn nil\n\t}\n\tresolved := make(map[string]SkillData, len(task.Agent.SkillRefs))\n\tmisses := make([]SkillRefData, 0)\n\tfor _, ref := range task.Agent.SkillRefs {\n\t\tref := ref\n\t\tvar bundle SkillData\n\t\tif err := d.skillCache.WithRefLock(task.WorkspaceID, ref, func() error {\n\t\t\tif cached, ok := d.skillCache.Load(task.WorkspaceID, ref); ok {\n\t\t\t\tbundle = cached\n\t\t\t\treturn nil\n\t\t\t}\n\t\t\tmisses = append(misses, ref)\n\t\t\treturn nil\n\t\t}); err != nil {\n\t\t\treturn fmt.Errorf(\"load skill bundle cache: %w\", err)\n\t\t}\n\t\tif bundle.ID != \"\" {\n\t\t\tresolved[skillRefKey(ref.Source, ref.ID)] = bundle\n\t\t}\n\t}\n\n\t// Resolve each missing bundle in its own request, caching it the moment it\n\t// arrives. The download is the slow part on jittery links, so fetching the\n\t// whole set in one atomic body read meant a single timeout discarded all\n\t// progress and the cache never converged — every dispatch re-downloaded\n\t// everything and timed out again. Per-skill, each download fits its own\n\t// size-scaled deadline and is persisted independently, so even a dispatch\n\t// that ultimately fails leaves the skills it did fetch cached for the next\n\t// one. (GitHub #4505 / MUL-3650)\n\tfor _, ref := range misses {\n\t\tstarted := time.Now()\n\t\tbundle, err := d.resolveSkillBundle(ctx, task, ref)\n\t\tif err != nil {","sourceCodeStart":5853,"sourceCodeEnd":5889,"githubUrl":"https://github.com/multica-ai/multica/blob/2c0912b6ec764b373d44eeea1e80f0d9f11ab417/server/internal/daemon/daemon.go#L5853-L5889","documentation":"Error \"load skill bundle cache: %w\" thrown in multica-ai/multica.","triggerScenarios":"Thrown at server/internal/daemon/daemon.go:5871 when the library encounters an invalid state.","commonSituations":"See trigger scenarios.","solutions":["Clear or rebuild the skill bundle cache and retry."],"exampleFix":null,"handlingStrategy":null,"validationCode":null,"typeGuard":null,"tryCatchPattern":null,"preventionTips":[],"tags":[],"backgroundTag":null,"analyzedSha":"2c0912b6ec764b373d44eeea1e80f0d9f11ab417","analyzedAt":"2026-08-15T13:25:18.241Z","schemaVersion":2},"datasetVersion":"2026-08-15T22:17:37.221Z"}