This roadmap reflects the current state of the public repo relative to the live private OpenClaw body.
It is ordered around real extraction work, not only presentation.
Already real in this repo:
- product README and trust layer
- boundary and privacy docs
- default agent pack skeleton
- Stage 1 context-intel extraction
- tests, fixtures, and CI for the extracted core
This stage proves:
- the repo is real
- the public lane is separate
- at least one useful public-safe slice already exists
Goal:
- replace the old marker view with a current extraction map derived from the live body
Deliverables:
- publicization matrix by surface
- richer public config example aligned to the real config shape
- onboarding flow that is honest about what is real vs roadmap
- runtime validation surface for the public core
Exit criteria:
EXTRACTION-MAP.mdclassifies every major lane ascopy,copy with templating,rewrite, orexcludeopenclaw.public.example.jsoncreflects the minimum installable public shape- onboarding no longer implies a setup surface that does not exist
Goal:
- package the minimum public-safe control plane needed for a real install
Current status:
- gateway templates are public
hyperclaw-materialize-packandhyperclaw-first-runare public- the public core can now be rendered over a clean target root
Deliverables:
- gateway and service templates
- public-safe validator entrypoints
- install scripts and first-run checks
- default pack enablement flow
Mostly expected status:
copy with templatingfor service units and validation wrappersrewritefor private-path-dependent helpers
Exit criteria:
- a user can configure and validate the public core without private files
- docs match the real install flow
Goal:
- move beyond Stage 1 context-intel into a real public-safe fabric layer
Current status:
- initial task, delegation, and watchdog schemas are now public
- bootstrap, validation, and summary wrappers are now public
- materialized installs now bootstrap public-safe runtime state by default
- richer dispatch and observability behavior is still pending
Deliverables:
- public-safe
query-fusionshell - dispatch, watchdog, and observability contracts
- synthetic fixtures for operational smoke tests
Mostly expected status:
copyfor schemas and public-safe fixturesrewritefor task capture, dispatch, watchdog, and live state handling
Exit criteria:
- the public repo exposes more than isolated tools
- memory and operational fabric are represented as real installable surfaces
Goal:
- package connector and voice surfaces as adapters, not as private residue
Current status:
- public connector env templates now exist for Telegram, HTTP hooks, Gmail watch, and calendar push
- live connector and voice automation remains a later adapter extraction
Deliverables:
- Telegram and WhatsApp template surfaces
- Gmail / Calendar / Drive hook templates
- voice/browser architecture docs and service templates
- public-safe RTC and voice adapter seams
Mostly expected status:
copy with templatingfor connector examples and service shellsrewritefor hook transforms, voice proxying, and provider-specific glue
Exit criteria:
- connector and voice lanes are documented as real optional surfaces
- no private tokens, allowlists, or session material are required in-repo
Goal:
- add richer optional lanes without bloating the public core
Current status:
- finance and legal now ship as explicit optional overlay boots
- Tier 5 repo-intel adapter extraction is still pending
Deliverables:
- GitNexus or equivalent repo-intel adapter contract
- optional finance and legal overlays
- optional specialist pack extensions for non-core lanes
- richer observability and mission-control style adapters
Exit criteria:
- optional lanes are useful but not required
- the base distro stays valuable without private overlays
Goal:
- reach a state where the repo can be presented as a real public distro without overstating maturity
Current status:
- boundary audit doc exists
- privacy scan now checks for private live config, runtime residue, and hard-coded private-root paths in installable surfaces
- repo visibility change is still blocked pending owner confirmation
Deliverables:
- owner privacy audit
- public-safety verification pass
- install and smoke-test proof
- docs pass across README, architecture, hosting, onboarding, and roadmap
Exit criteria:
- packaged reality matches public claims
- private-only surfaces are clearly excluded
- repo is credible even without grant framing
- copying the private body wholesale
- pretending feature parity with the live system
- making optional adapters mandatory
- claiming installability before the install surface is real