Code
Hub
Workspaces
Following
Trending
Connect
MCP
copy
Create free account
hub
/
github.com/GizClaw/flowcraft
/ types & classes
Types & classes
2,684 in github.com/GizClaw/flowcraft
⨍
Functions
13,468
◇
Types & classes
2,684
↓ 20 callers
FuncType
optionFunc
func(*ltmConfig)
sdk/recall/pipeline/factory.go:22
↓ 7 callers
TypeAlias
Pattern
Pattern is a Subject matcher using NATS-style wildcards: - matches exactly one segment > matches one or more trailing segments (must be the last seg
sdk/event/subject.go:31
↓ 7 callers
TypeAlias
Subject
Subject is a dot-delimited routing key, e.g.: graph.run.r1.start graph.run.r1.node.n1.complete kanban.board.b1.update Segments are separated by '
sdk/event/subject.go:17
↓ 5 callers
TypeAlias
RetrieveMode
RetrieveMode is one source for MultiRetrieve / Retrieve. The string value of each mode MUST match the corresponding [retrieval.LaneKey] constant — th
memory/retrieval/pipeline/stages_recall.go:20
↓ 5 callers
TypeAlias
RetrieveMode
RetrieveMode is one source for MultiRetrieve / Retrieve. The string value of each mode MUST match the corresponding [retrieval.LaneKey] constant — th
sdk/retrieval/pipeline/stages_recall.go:20
↓ 5 callers
FuncType
deciderFunc
helper: deciderFunc adapts a closure into agent.Decider.
sdk/agent/run_test.go:746
↓ 5 callers
FuncType
stubProvider
stubProvider is a tiny adapter used to inject canned responses into a Multi without writing a full type for each test.
cmd/vesseld/secrets/secrets_test.go:377
↓ 4 callers
TypeAlias
Category
Category classifies a long-term memory entry.
sdk/recall/entry.go:9
↓ 3 callers
FuncType
BoardSeederFunc
BoardSeederFunc is the function-typed adapter for BoardSeeder. Useful when the seed logic is a single closure over a transcript loader or retriever:
sdk/agent/seeder.go:52
↓ 3 callers
TypeAlias
EventType
EventType is a publish Claw stream event discriminator.
sdkx/claw/roundtrip.go:23
↓ 3 callers
FuncType
Func
func(audio.Frame) audio.Frame
voice/preprocess/preprocess.go:9
↓ 3 callers
Interface
LLM
LLM is the unified interface for language model interactions.
sdk/llm/llm.go:118
↓ 2 callers
TypeAlias
ErrorCode
voice/error.go:10
↓ 2 callers
FuncType
HookFunc
func(TurnMetrics)
voice/metrics/metrics.go:35
↓ 2 callers
FuncType
MetricsSinkFunc
MetricsSinkFunc adapts a function to MetricsSink.
memory/recall/ops/types.go:36
↓ 2 callers
TypeAlias
RevisionKind
RevisionKind classifies how a write relates to prior facts. - RevisionSupersede: default merge_key conflict path (close prior). - RevisionFork: paral
memory/recall/internal/domain/revision.go:11
↓ 2 callers
TypeAlias
State
State is the per-task mutable "world". Concretely it's a JSON-like nested map (orders, products, customers, …) that tools read from and mutate. We use
eval/taubench/taubench.go:93
↓ 2 callers
TypeAlias
VideoID
VideoID represents a YouTube video ID.
sdkx/extract/youtube/detect.go:9
↓ 2 callers
FuncType
call
func(map[string]any) error
sdk/script/bindings/llm_marshal_test.go:703
↓ 2 callers
TypeAlias
staticEnumerator
memory/recall/ops/runner_test.go:578
↓ 1 callers
TypeAlias
BackpressurePolicy
BackpressurePolicy controls what happens when a subscription's buffer is already full at Publish time.
sdk/event/observer.go:9
↓ 1 callers
TypeAlias
CardStatus
CardStatus represents the lifecycle state of a Card on the kanban board.
sdk/kanban/card.go:6
↓ 1 callers
TypeAlias
CollapseStrategy
CollapseStrategy names the chunks→docID aggregation function used by runLane. See Options.CollapseStrategy for the full rationale.
eval/beir/beir.go:254
↓ 1 callers
TypeAlias
DropReason
DropReason explains why an envelope was discarded for a particular subscription.
sdk/event/observer.go:63
↓ 1 callers
FuncType
EngineBuilder
EngineBuilder is the resolved engine factory closure, already bound to its ref string + config map. The fleet calls it from inside its EngineFactory c
cmd/vesseld/resolver/plan.go:198
↓ 1 callers
FuncType
EngineFactory
EngineFactory builds an [engine.Engine] for the given Agent entry, receiving the resolved dependencies the Captain has hydrated. The factory is called
vessel/option.go:31
↓ 1 callers
TypeAlias
ErrorKind
ErrorKind enumerates the errdefs categories scripts may set on signal.error({ kind, message }). Only categories that can plausibly originate from scri
sdk/script/signal.go:20
↓ 1 callers
TypeAlias
JobID
----------------------------------------------------------------------------- Public types -----------------------------------------------------------
sdk/recall/jobs.go:24
↓ 1 callers
TypeAlias
MemoryFactKind
MemoryFactKind is a public semantic memory fact kind accepted by AppendMemoryFact.
sdkx/claw/memory_fact.go:20
↓ 1 callers
TypeAlias
MergeStrategy
MergeStrategy defines how parallel branch results are merged.
sdk/graph/runner/internal/executor/executor.go:115
↓ 1 callers
FuncType
ObserverFunc
func(Report)
voice/provider/fallback.go:42
↓ 1 callers
TypeAlias
ResolveOp
ResolveOp is the action an [UpdateResolver] requests for one existing memory in response to a new fact.
sdk/recall/resolver.go:15
↓ 1 callers
TypeAlias
Strategy
Strategy enumerates the three loading modes the bench compares.
eval/history/eval.go:22
↓ 1 callers
TypeAlias
SubscriptionID
SubscriptionID identifies a subscription within a single Bus instance. It is opaque to consumers; bus implementations choose the format.
sdk/event/observer.go:5
↓ 1 callers
TypeAlias
WriteMode
WriteMode controls SaveRequest semantics. Zero value = synchronous (preserves current behaviour). WriteModeAsyncSemantic stores raw episodes synchrono
memory/recall/internal/domain/write_mode.go:6
↓ 1 callers
TypeAlias
appendAIDeltaMsg
examples/voice-pipeline/tui.go:23
↓ 1 callers
Struct
chatLine
examples/voice-pipeline/tui.go:14
↓ 1 callers
TypeAlias
errSentinel
memory/recall/internal/ingest/view.go:29
↓ 1 callers
TypeAlias
errSentinel
sdk/telemetry/log_test.go:370
↓ 1 callers
TypeAlias
fakeRunnerErr
sdk/script/bindings/bridge_shell_test.go:35
↓ 1 callers
TypeAlias
multiObserver
sdk/agent/observer.go:153
↓ 1 callers
TypeAlias
stringError
cmd/vesseld/fleet/sessionstore_test.go:176
↓ 1 callers
FuncType
stubFn
func() (FinalizeDecision, error)
sdk/agent/decider_test.go:106
↓ 1 callers
TypeAlias
updatePartialMsg
examples/voice-pipeline/tui.go:21
Interface
AckSubscription
AckSubscription is the optional capability a Subscription may expose when its parent Bus implements at-least-once delivery (typically cross-process bu
sdk/event/bus.go:138
Interface
AcousticEchoCanceller
AcousticEchoCanceller optionally preprocesses microphone capture using a playback reference frame before speech detection. Production implementations
voice/detect/echo.go:17
Struct
ActivatedLens
ActivatedLens is one row in PlanDetail.ActivatedLenses. Lives in shared.go (cross-detail vocabulary, per restructure §4).
memory/recall/internal/domain/diagnostic/shared.go:204
FuncType
ActorFn
ActorFn extracts an actor label from ctx (optional).
memory/retrieval/journal/wrap.go:11
FuncType
ActorFn
ActorFn extracts an actor label from ctx (optional).
sdk/retrieval/journal/wrap.go:11
Struct
AdditiveExtractor
AdditiveExtractor is the default single-pass extractor. When LLM is nil, Extract returns an empty slice; callers should treat Add as the only ingesti
sdk/recall/extractor.go:511
Struct
Agent
Agent describes one agent inside a vessel. The Captain assembles the corresponding sdk/agent.Agent value at New time using these fields plus the depen
vessel/spec/spec.go:85
Struct
Agent
Agent is the wire form of a single vessel agent. Only the fields the daemon needs to differentiate at startup live here; rich engine config (system pr
cmd/vesseld/apispec/v1alpha1/agent.go:13
Struct
Agent
Agent is the application-layer description of one logical agent. It is a plain data struct, NOT a "live object" that knows how to run itself: executio
sdk/agent/agent.go:11
Struct
AgentCapabilities
AgentCapabilities declares which optional A2A features the agent supports. JSON keys exactly match the A2A spec — note the plural PushNotifications an
sdk/agent/agent.go:122
Struct
AgentCard
AgentCard mirrors the agent.AgentCard wire fields the resolver can populate. We do NOT reuse the SDK type directly so wire changes here are independen
cmd/vesseld/apispec/v1alpha1/agent.go:71
Struct
AgentCard
AgentCard describes an agent's capabilities for discovery. Field names and JSON tags are a *proper subset* of the A2A AgentCard specification: every f
sdk/agent/agent.go:66
Struct
AgentConfig
AgentConfig configures the single agent owned by a Claw.
sdkx/claw/agent.go:19
Struct
AgentEngine
AgentEngine is the engine ref + opaque config map. The factory registered under Ref decides what the Config keys mean. We use map[string]any rather th
cmd/vesseld/apispec/v1alpha1/agent.go:83
Interface
AgentExecutor
AgentExecutor is the interface for executing tasks on target agents. WorkflowRunner implements this by creating/reusing AgentActors.
sdk/kanban/executor.go:7
Struct
AgentSpec
vessel/assembly/manifest.go:118
Struct
AgentSpec
AgentSpec carries the per-agent configuration.
cmd/vesseld/apispec/v1alpha1/agent.go:20
FuncType
AgentValidator
AgentValidator checks whether a target agent ID is valid before submission. Return nil if valid; return an error (ideally listing available IDs) if no
sdk/kanban/kanban.go:131
Interface
AliasResolver
AliasResolver canonicalises a single surface mention within a scope. Implementations stay pure (no I/O); they are consulted once per mention per Inges
memory/recall/internal/port/ingest.go:97
Struct
AnswerContext
AnswerContext is an answer-ready prompt body produced by a backend from its native recall results. Backends with structured memory should prefer this
eval/locomo/runners/runner.go:55
Struct
AnswerContextItem
AnswerContextItem is one rendered memory item passed to an answer stage.
memory/recall/diagnostics/recall.go:194
Interface
AnswerContextRecaller
AnswerContextRecaller lets a backend keep its native recall result shape for answer rendering. The returned artifacts are for diagnostics and report d
eval/locomo/runners/runner.go:104
Interface
AnswerContextStageAuditor
AnswerContextStageAuditor combines structured answer rendering with recall stage diagnostics.
eval/locomo/runners/runner.go:110
Struct
AnswerQuestion
AnswerQuestion is the small part of a benchmark question a backend needs to render its own answer context. It deliberately avoids depending on the eva
eval/locomo/runners/runner.go:47
Struct
AnswerReplayArtifact
eval/locomo/answer_replay.go:34
Struct
AnswerReplayOutcome
eval/locomo/answer_replay.go:46
Struct
AnswerReplayRecord
AnswerReplayRecord is the diagnostic JSONL shape emitted by --dump-answer-replay. It captures the exact answer input and scored output so verifier/aud
eval/locomo/answer_replay.go:16
Struct
Append
Append writes the resolved facts to the canonical store. The compensator deletes whatever was appended on any downstream failure. When the downstream
memory/recall/internal/pipeline/write/stages/append.go:25
Struct
AppendDetail
AppendDetail —— write/append stage.
memory/recall/internal/domain/diagnostic/write_detail.go:64
Struct
AppendEpisode
AppendEpisode writes the raw KindEpisode facts produced by build_episode to the canonical store. It is separate from Append so the runner / trace clea
memory/recall/internal/pipeline/write/stages/append_episode.go:20
Struct
AppendMemoryFactResult
AppendMemoryFactResult reports the canonical memory facts written by AppendMemoryFact.
sdkx/claw/memory_fact.go:86
Interface
Appender
Appender is an optional interface. If the underlying store supports native append (e.g. Alibaba OSS AppendObject), it should implement this. The works
sdkx/workspace/objstore/store.go:50
Struct
ApplyFeedback
ApplyFeedback is the single stage of the feedback pipeline. Run order: 1. Validate FactID + at least one positive delta. 2. store.UpdateFeedback app
memory/recall/internal/pipeline/feedback/stages/apply_feedback.go:36
Struct
ArchiveConfig
ArchiveConfig controls message archiving behavior.
memory/history/compactor.go:856
Struct
ArchiveConfig
ArchiveConfig controls message archiving behavior.
sdk/history/compactor.go:856
Struct
ArchiveManifest
ArchiveManifest tracks archived message segments.
memory/history/archive.go:28
Struct
ArchiveManifest
ArchiveManifest tracks archived message segments.
sdk/history/archive.go:28
Struct
ArchiveResult
ArchiveResult holds the result of an archive operation.
memory/history/archive.go:43
Struct
ArchiveResult
ArchiveResult holds the result of an archive operation.
sdk/history/archive.go:43
Struct
ArchiveSegment
ArchiveSegment describes a single archived file.
memory/history/archive.go:34
Struct
ArchiveSegment
ArchiveSegment describes a single archived file.
sdk/history/archive.go:34
Struct
Artifact
Artifact is a named bundle of typed parts produced during a run (e.g. "summary", "tool_output_image"). Engines that write artifacts store them in a bo
sdk/agent/result.go:52
Struct
Assembly
Assembly is the resource graph produced by Build.
vessel/assembly/types.go:19
Struct
AsyncSemanticClaimOptions
AsyncSemanticClaimOptions controls Claim batching and tenancy filters. Zero Max defaults to no jobs; callers should set Max explicitly. ProcessAsyncSe
memory/recall/internal/port/async_semantic.go:78
Struct
AsyncSemanticFailure
AsyncSemanticFailure is the worker's failure report. ErrClass drives retry vs dead-letter routing in the queue backend.
memory/recall/internal/port/async_semantic.go:134
Struct
AsyncSemanticJob
AsyncSemanticJob is one durable semantic extraction work item.
memory/recall/internal/port/async_semantic.go:87
Struct
AsyncSemanticProcessDetail
AsyncSemanticProcessDetail —— async semantic worker completion. Emitted once per processed job so operators can join sync SaveTrace stages via AsyncRe
memory/recall/internal/domain/diagnostic/write_detail.go:247
Struct
AsyncSemanticProcessOptions
AsyncSemanticProcessOptions controls one ProcessAsyncSemantic batch.
memory/recall/async_semantic.go:26
Struct
AsyncSemanticProcessResult
AsyncSemanticProcessResult summarizes one drain pass.
memory/recall/async_semantic.go:39
Interface
AsyncSemanticProcessor
AsyncSemanticProcessor is the caller-driven drain entry for queued semantic extraction. It is intentionally separate from Memory so the main facade do
memory/recall/async_semantic.go:21
Interface
AsyncSemanticQueue
AsyncSemanticQueue is a durable outbox for semantic extraction work items produced by Memory.Save(WriteModeAsyncSemantic). Implementations MUST: - Ma
memory/recall/internal/port/async_semantic.go:23
FuncType
AsyncSemanticQueueFactory
AsyncSemanticQueueFactory returns a fresh, empty AsyncSemanticQueue for one subtest.
memory/recall/recalltest/async_semantic.go:12
Struct
AsyncSemanticQueueHealth
AsyncSemanticQueueHealth is the operator-facing queue snapshot used by health dashboards.
memory/recall/diagnostics/async_queue.go:9
Interface
AsyncSemanticQueueObserver
AsyncSemanticQueueObserver exposes queue health when Memory is wired with WithAsyncSemanticQueue and the backend implements Stats.
memory/recall/async_semantic.go:59
Struct
AsyncSemanticReceipt
AsyncSemanticReceipt is the return value of Enqueue and the wire contract observable to Memory.Save callers via SaveResult.
memory/recall/internal/port/async_semantic.go:115
Struct
AsyncSemanticReconcileOptions
AsyncSemanticReconcileOptions controls canonical async semantic recovery.
memory/recall/reconcile.go:40
next →
1–100 of 2,684, ranked by callers