modulePackage manifestv2.1.4

module.agent-fleet

Agent fleet management module — fleet, tree, detail, tool registry, and run console surfaces.

@lssm-tech/module.agent-fleet
Source, release and verification context

Source snapshot bcc79fcc · Declared package version 2.1.4. This page combines the checked-in package guide and manifest; commands and runtime behavior were not executed in this review.

Package source ↗ Manifest ↗
servernode
EXPLORE THE PACKAGE

See where this package fits.

Follow a declared connection, inspect an import, then open the source guide when you need the detail.

THIS PACKAGEmodule.agent-fleetplatform

Arrows point from a package to its direct dependency. Each count is derived from the captured manifests.

14 déclarations · les 6 premières affichéesSELECT TO INSPECT

@lssm-tech/module.agent-fleet

Agent fleet, tool-registry, and governed execution surfaces. The package also exports two distinct installable Feature Hub profiles:

  • ./agent-system for local fleet, tool, memory, lease, and audit adapters.
  • ./agent-network for cross-system delegation and federation.

Federation verification uses RFC 8785 canonical UTF-8 bytes with the signature omitted, intersects every signed/requested/current/exposed scope and ceiling, and fails closed through injected trust, key, replay, revocation, target-local policy, and audit ports. Replay is consumed before target-local authorization; over-scoped requests do not reach local policy. A local allow carrying unresolved obligations remains denied, so a verification result cannot silently discard target controls.

External Markdown migration preview

previewAgentMarkdownMigration validates the restricted archive and exact agent scope, then projects explicit target mappings as new, identical or conflicting. Memory requires review by its owning PersonalOS/CompanyOS service; skill content requires independent evaluation. The projection contains hashes and references, not imported text. It neither writes files nor installs or activates content. Product import UI, durable owner proposals, schedule/channel translation and live vendor migration qualification remain separate work.

proposeAgentMarkdownMemory translates an explicitly selected memory file into the existing AgentMemoryProposalArtifactV1. It binds source/archive/content digests, exact scope, host-selected owner and policy references, and an optional correction target/revision. Imported contents are not embedded in the proposal or interpreted as authority. The host must resolve contentRef in restricted storage; PersonalOS or CompanyOS must verify consent, retention and authority before persistence.

submitAgentMemoryImportProposal composes the existing agent protocol artifact store with host-provided submission authorization and restricted-content verification. It appends at revision zero, verifies exact readback, and reconciles lost acknowledgements without repeating the append in the same call. Unverified writes remain indeterminate. This persists review candidates only; memory-owner promotion, consent and runtime activation stay separate. Production hosts must provide the durable scoped adapter and current principal authorization.

AgentMigrationReviewPanel is a controlled reusable presentation for console or embedded hosts. Pass a validated preview, localized labels, selected target refs, pending/error state and host permission projection. It allows selection only for new mappings and sends the archive digest plus selected target refs to the host. The host must revalidate all input and call the authorized submission service; client-side canSubmit never grants permission. Conflicts require mapping resolution outside this panel. Product-route wiring and browser/visual qualification are pending.

Migration submission has one monotonic I/O budget (default ten seconds, configurable from one millisecond to sixty seconds). All authorization, content verification, artifact reads/writes and record validation share it. Preflight timeout starts no write; an unsettled append remains indeterminate. The generic store interface cannot forcibly cancel a provider request, so explicit retries must reconcile the same immutable proposal identity rather than creating another proposal.

Submission authorizes the actor before restricted-content lookup and rechecks current authority immediately before a new artifact append. Revocation during lookup therefore blocks the write; it does not rely solely on an earlier decision.

executeMemoryImportOperation is the handler adapter for agent.memory.import.submit. A host-provided Runtime Node verifier must return current signed-operation bindings matching the exact request, actor, scope, work-order, archive/proposal digests and idempotency key. Restricted artifacts are loaded only after that check; authority is rechecked before submission. The adapter never creates a signature, claim, fence or promotion decision. Production verifier composition, handler registration and HTTP/UI route wiring remain required.

createMemoryImportRuntimeVerifier composes the canonical V3 order validator, Connect dispatch verifier and durable operation reader. The signed payload binds the actor, agent scope and complete request; the operation must retain the same running claim/attempt/fence and request digest after verification. Trusted time is rechecked for expiry. Production key verification still belongs to Connect; this adapter neither implements a signing authority nor accepts bare digest equality as signature proof. Hosts must provide the real trusted readers and verifier.

createMemoryImportOperationRegistry returns a canonical operation registry with the import command and handler bound. Normal registry input/output and permission checks apply. The handler additionally binds the authenticated context user and tenant to the verified Runtime Node actor/scope, preventing a transport caller from selecting another identity. Hosts still supply the production verifier and store adapters and must register the registry with their authenticated transport.

Memory import registry callers must supply a server-authenticated workspaceId alongside actor and physical tenant. Both the request and verified Runtime Node operation must match that workspace; a tenant-wide session alone cannot import.

Founder operations workflows (./core)

These additive surfaces are headless and provider-neutral. Hosts inject durable stores, provider control, notification, authority and Work-creation ports. Nothing here executes consequential effects on its own.

  • coding-supervision-v1 — SupervisedCodingSessionService binds a coding agent run to a canonical Work item (UUIDv7), repository/worktree/branch, path leases and an exact task digest. Questions, reviews and CI waits are persisted before notification; answers must match the wait id, expected revision and current owner epoch and pass current authority. Delivery is pending → dispatching → delivered; after a restart sweep() dispatches never-sent answers once and reconciles in-flight ones by provider lookup. Uncertain provider, delivery and cancellation outcomes stay explicit and are escalated, never retried blindly. Wait deadlines, wall-clock/compute/cost budgets and stalls escalate once per trigger as a FounderProposalV1 (proposed Work item); onExhausted: 'cancel_and_escalate' additionally requests cancellation. Completion needs every declared artifact; acceptance needs the required verification receipts. External sessions are unqualified until an identity proof is supplied.
  • Proactive proposals (FW-013) — ProactiveProposalService accepts explained candidates (trigger, evidence, expected value, affected scope, next action), suppresses duplicates and repeatedly dismissed rules, keeps paused or archived scopes quiet unless an obligation remains, and defers routine items, quiet hours and rate-limited items to the digest. Critical blockers may bypass quiet hours and the rate limit per policy. Accepting returns a Work draft only.
  • workflow-improvement-v1 — WorkflowImprovementService and createWorkflowImprovementAdapter route knowledge corrections to Company Intelligence and send procedure feedback through GovernedAiImprovementRuntimeV2 (@lssm-tech/lib.ai-agent/improvement-loop) with a frozen success criterion, replay plus independent held-out evaluation, canary observation and rollback. Failed evaluations are rejected; promotion requires an explicit human approval; self-granted authority and model-route drift fail closed.
  • usecase-proposal-v1 — reviewUseCaseProposals turns recurring observed work into at most maxProposals proposals with past examples, frequency, benefit, needed access (granted: false), operating cost, smallest trial, stop condition and acceptance hypothesis, at the agreed cadence. Nothing is activated.

Reference stores (InMemorySupervisedCodingSessionStore, InMemoryProactiveProposalLedger, InMemoryWorkflowImprovementBriefStore) are deterministic fixtures, not production persistence. For production, pass supervisedCodingSessionCodecV1, proactiveLedgerRecordCodecV1 and workflowImprovementBriefCodecV1 to the matching Postgres stores in @lssm-tech/integration.agent-durable-store/postgres. The codecs validate every persisted payload. One shared contract suite (src/core/store-contracts/) runs against both the in-memory and the PGlite-backed Postgres implementations.

createCommunicationQuietHoursInterruptionPort binds proactive quiet hours to CommunicationOS isWithinCommunicationQuietHoursV1. A missing preference means there is no quiet window. A preference that cannot be read fails toward quiet.

Declared capabilities.

Module Agent Fleet

Agent fleet management module — fleet, tree, detail, tool registry, and run console surfaces.

Use it to

  • Adopt agent fleet management module — fleet, tree, detail, tool registry, and run console surfaces

Requires

  • TypeScript

Choose another boundary when

  • The @lssm-tech/module.agent-fleet package constraints do not match the target runtime or platform.

@lssm-tech/module.agent-fleet

@lssm-tech/module.agent-fleet/agent-fleet.capability

@lssm-tech/module.agent-fleet/agent-fleet.feature

@lssm-tech/module.agent-fleet/core

@lssm-tech/module.agent-fleet/data-views

@lssm-tech/module.agent-fleet/hubs

@lssm-tech/module.agent-fleet/hubs/agent-network

@lssm-tech/module.agent-fleet/hubs/agent-system

@lssm-tech/module.agent-fleet/hubs/federation-ports

@lssm-tech/module.agent-fleet/hubs/federation-verification

@lssm-tech/module.agent-fleet/hubs/presentation

@lssm-tech/module.agent-fleet/presentation

Imports and export paths.

12 entries

These are package-manifest entry points. Named symbols listed above come from the capability catalogue; this is not an exhaustive TypeScript symbol reference.

12 imports

Dependencies and related packages.

Runtime dependencies 14
react
catalog:
Peer dependencies 1
react
catalog:
Development dependencies 7
@electric-sql/pglite
catalog:data
@types/react
catalog:types
typescript
catalog:

Package scripts.

Commands declared by this package. Review its setup requirements before running them; no command here was executed by this review.

publish:pkgbun publish --tolerate-republish --ignore-scripts --verbose
publish:pkg:canarybun publish:pkg --tag canary
buildcontractspec-bun-build build
build:bundlecontractspec-bun-build transpile
build:typescontractspec-bun-build types
devcontractspec-bun-build dev
cleanrimraf dist .turbo
lintbun lint:fix
lint:fixnode ../../../../scripts/biome.cjs check --write --unsafe --only=nursery/useSortedClasses . && node ../../../../scripts/biome.cjs check --write .
lint:checknode ../../../../scripts/biome.cjs check .
testbun test
prebuildcontractspec-bun-build prebuild
typechecktsgo --noEmit
Votre vue

Make room for your context.

Shape the presentation around this moment. You can change it whenever you need.

Votre contexte

The different parts of you

Choose a starting point

Un mélange utile de contexte, de visuels et de guidage discret.

Tune each dimension
Each dimension can change independently. Product permissions and available actions stay unchanged.

Support the way you begin and return

Tune your support

Confort de lecture

Choisir votre langue

La langue de l’interface change uniquement le texte. Les marchés, les prix et les parcours conservent leurs propres paramètres.

Couverture des traductions

Une partie du contenu est affichée en anglais.

Navigation151
Graphe616
Catégories378
Portfolio833
Exemples2 035
Aperçu de la source1 542

D’autres traductions seront ajoutées à mesure que le contenu sera relu.

Cette préférence est enregistrée dans ce navigateur.

Rechercher sur le site

Des points de départ utiles

Aucun résultat. Essayez une expression plus courte, un nom d’opération ou un autre périmètre de recherche.

UN PEU DE PLACE POUR EXPLORER

Trouvez quelque chose à explorer.

Ouvrez un aperçu détaillé sur cette page, essayez une autre perspective ou suivez le détail vers sa propre page.

Fonctionnement des aperçus et des commandes
01

Regarder avant de partir

Survolez un lien ou sélectionnez-le avec Tab pour voir un aperçu. Cliquer sur le lien d’origine ou appuyer sur Entrée ouvre toujours sa destination.

02

Rester dans le contexte

Entrez dans l’aperçu pour explorer ses sections. Depuis un lien sélectionné, utilisez Alt + ↓ pour y entrer au clavier. Échap le ferme.

03

Ouvrir les possibilités

Les lignes dépliables, les choix de perspective et les commandes mises en évidence révèlent davantage. Votre vue adapte le confort de lecture et le contexte exploré.

Un bouton séparé pour voir de plus près

Affichez de petits boutons d’aperçu à côté des liens quand l’espace le permet. Touchez l’œil pour examiner ou le lien pour y aller directement. Ce choix reste dans cet onglet.

Aucun historique de navigation n’est collecté. La réduction des animations suit votre appareil et Votre vue.