AK-Bench / Knowledge base
Documentation
Explore the project, learn the tools, and find the details behind each benchmark.
53documents
Project
1 documentHandbook
18 documentsAgent adaptersdocs/AGENT_ADAPTERS.mdAgent integration — spike results (Phase 0)docs/AGENT_INTEGRATION.mdCLIdocs/CLI.mdDatasets, tasks and workspaces (Phase 3)docs/DATASETS.mdDescriptor system (Phase 2)docs/DESCRIPTOR_SYSTEM.mdDevelopment environmentdocs/DEVELOPMENT.mdEvaluator Enginedocs/EVALUATORS.mdExperiments and planningdocs/EXPERIMENTS.mdExportdocs/EXPORT.mdKnowledge Replaydocs/KNOWLEDGE_REPLAY.mdMetrics and Observabilitydocs/METRICS.mdPlugin Configuration and Observabilitydocs/PLUGINS.mdReportingdocs/REPORTING.mdReproducibilitydocs/REPRODUCIBILITY.mdLLM Router integration — spike results (Phase 0)docs/ROUTER_INTEGRATION.mdRuntime Managerdocs/RUNTIME_MANAGER.mdRun Orchestratordocs/RUN_ORCHESTRATOR.mdStoragedocs/STORAGE.md
Datasets
31 documentsReadmedatasets/README.mdbookshopdatasets/_sources/bookshop/README.mdcadencedatasets/_sources/cadence/README.mdDepweaverdatasets/_sources/depweaver/README.mdResolution and repair contractdatasets/_sources/depweaver/SPEC.mdfreightdatasets/_sources/freight/README.mdledgerlitedatasets/_sources/ledgerlite/README.mdnotesapidatasets/_sources/notesapi/README.mdstockpiledatasets/_sources/stockpile/README.mdStreamforgedatasets/_sources/streamforge/README.mdtextkitdatasets/_sources/textkit/README.mdTypeweavedatasets/_sources/typeweave/README.mdType inference implementation contractdatasets/_sources/typeweave/SPEC.mdWALQueuedatasets/_sources/walqueue/README.mdCancelling a confirmed order leaves the stock reserveddatasets/bug_fix/bookshop-stock-002/description.mdThe five percent tier is lost at exactly one hundred dollarsdatasets/bug_fix/bookshop-tiers-001/description.mdA worker with a stale epoch can take a live lease backdatasets/bug_fix/cadence-takeover-001/description.mdRepair the resolver's branch and context protocoldatasets/bug_fix/depweaver-backtracking-001/description.mdA conversion answers with the oldest rate in forcedatasets/bug_fix/ledgerlite-rates-001/description.mdA write-off destroys the reserved part of an expired lotdatasets/bug_fix/stockpile-writeoff-001/description.mdRepair durable queue recoverydatasets/bug_fix/walqueue-recovery-001/description.mdWord counting must ignore repeated whitespacedatasets/bug_fix/wordcount-001/description.mdImplement `plan_catch_up` from its specificationdatasets/new_feature/cadence-catchup-001/description.mdImplement `shop` from its specificationdatasets/new_feature/freight-shop-001/description.mdImplement `reconcile` from its specificationdatasets/new_feature/ledgerlite-reconcile-001/description.mdImplement `search_notes` from its specificationdatasets/new_feature/notes-search-001/description.mdImplement `merge` from its specificationdatasets/new_feature/notes-sync-001/description.mdImplement `slugify` from the specificationdatasets/new_feature/slugify-001/description.mdImplement `allocate` from its specificationdatasets/new_feature/stockpile-fefo-001/description.mdImplement a restartable event-time enginedatasets/new_feature/streamforge-windows-001/description.mdComplete the type checker of an existing compilerdatasets/new_feature/typeweave-inference-001/description.md