Skip to content

Latest commit

 

History

567 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

WORK-LAB

DTALEX66/WORK-LAB is the canonical monorepo control plane for the user's client-neutral AI workflow global-configuration layer. It manages the highest user-level capability layer — Rules, Skills, plugins/MCP declarations, portable Memory, Capabilities and workflow policy — as one canonical source, then adapts it into native projections for each current client. It is not an agent runtime, a product platform, or a fourth product.

Current positioning (V2 converged): the two canonical modules are packages/client-neutral-core (task/telemetry ledger, sidecar, adapters, delivery gates) and apps/observer (strictly read-only projection). The supporting source surfaces services/, integrations/, config/ and apps/token-monitor/ serve those modules and are not separate module roots. Managed clients: Hermes · Codex · DSH · GitHub · Open Design · OpenHuman, plus any future AI software through the same Adapter contract; CC Switch is LEGACY_OBSERVE (observe-only, no active writes). Frozen history lives under docs/history/ (single retrieval anchor; see WL_INHERITANCE_MATRIX.json, MiniGame = FOREIGN_HISTORICAL). See docs/decisions/PROJECT_POSITIONING.md.

Neutrality (unbound, unlocked)

  • Client-neutral: one canonical source → per-client native projection, not byte-identical copies.
  • Unbound: the current clients (Hermes · Codex · DSH · GitHub · Open Design · OpenHuman; CC Switch is LEGACY_OBSERVE, observe-only) are the current adapters, not permanent dependencies; future AI software plugs into the same contract. DSH is a replaceable agent runtime (temporary executor in the model control-plane taskpack), not a Hermes replacement; see docs/decisions/PROJECT_POSITIONING.md.
  • Unlocked: core schemas use stable IDs and capability discovery — never hard-coded programs, model IDs, versions, ports, or install paths.

Workspace mission

WORK-LAB turns a request that spans planning, execution, review and delivery into an auditable task-pack loop:

request / task pack
  -> ownership + allowed paths + data boundary
  -> one writer with read-only parallel audits
  -> RED -> GREEN -> targeted/module/aggregate gates
  -> exact tree review + recovery evidence
  -> explicit commit/push/release approval

The global workflow surface is repository-controlled portable source under packages/client-neutral-core. Hermes Home, credentials, sessions, cron metadata, provider routes and live caches remain global platform state; they are never absorbed into this repository. Project task data and generated evidence stay under the current Git root's ignored .project-local/ boundary.

Canonical active modules

The machine module model (.project/governance/projects.json) defines exactly two canonical modules:

  • packages/client-neutral-core (module id workflow-assistance) — global, client-neutral workflow governance, task/telemetry ledgers, sidecar, adapters and delivery gates
  • apps/observer (module id work-lab-observer) — strictly read-only derived observation, projections and evidence reports

Active implementation / supporting source surfaces that serve those modules (not separate module roots): services/ (orchestration, policy, execution-federation, receipts, session-federation, task-governance, radar, security and cleanup), integrations/, config/ and apps/token-monitor/.

Frozen history lives under docs/history/ (single retrieval anchor, incl. converged report history from 90-archive/ via MOVE-005; 90-archive/ retains only its BOUNDARY.md gate marker). MiniGame and pre-cutover r4-era material remain archive evidence, not active modules.

The repository preserves source history and keeps module implementations under their canonical prefixes. Root governance owns stable contracts, task/evidence formats, CI, release policy and cross-module ownership; it does not duplicate module implementations.

Task-pack responsibility

All substantial work is represented by a reviewed task pack under taskpacks/current/. A task pack must identify the writer, read-only reviewers, allowed/forbidden roots, input evidence, completion contract, verification commands, rollback handle and external-mutation approval. Missing evidence, dirty ownership, boundary violations or incomplete required jobs fail closed.

Safety boundaries

  • One Git root; module rules may narrow root rules but never weaken them.
  • .project-local/ holds ignored runtime + evidence data; frozen history lives under docs/history/ (single retrieval anchor — 90-archive/ converged into it, retaining only its BOUNDARY.md gate marker). Durable handoffs may be retained only when explicitly covered by a recovery contract.
  • Secrets, credentials, prompt/response bodies and private browser data never enter evidence.
  • Project artifacts never spill outside the project Git root; any spill is traceable, locatable, cleanable and migratable (see .project/governance/project-data-boundary.json).
  • The E: and F: data volumes are protected: any access requires explicit per-path, per-operation user authorization.
  • External mutation, active-path switching and release actions require explicit approval.
  • Windows paths are checked case-insensitively before release.

See docs/decisions/PROJECT_POSITIONING.md, .project/governance/projects.json, module AGENTS.md files, and taskpacks/current/WORK-LAB-UNIFIED-PRODUCT-CONVERGENCE-TASKPACK-20260918.md for the current positioning and migration record.

About

Client-neutral AI-agent control plane: user global config overlay for Hermes/Codex/DSH/GitHub/Open Design/OpenHuman via one adapter contract (CC Switch observe-only). Modules: client-neutral-core (task/telemetry ledger, sidecar), services (orchestration/policy/receipts), read-only observer. Authority chain + exact-SHA CI gates.

Topics

Resources

Security policy

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages