[pcvantol]

Same LANE_4 assignment

This is the existing PR #35 for L4-WORKSPACE-SERVER-READONLY-V1-20261001 (Forge register #208, revision 1). The owner-selected entry reconciliation and the first own read-only Server outcome are now on the same branch. No replacement PR or assignment was created. Forge four-lane proposal #209 remains a separate owning change.

Delivered candidate

• Own private-root Server with stable opaque instance identity/token, process writer lock, restart-safe own catalogue and two-instance isolation.
• Loopback-only versioned HTTP with bearer authentication, required pinned instance identity, strict read-only methods, origin/path negatives, source-state semantics and OpenAPI/Postman route parity.
• Thin local Server CLI and browser Client using actual Server HTTP responses; the browser distinguishes UNCONFIGURED, EMPTY, PARTIAL, STALE, UNAUTHORIZED, WRONG INSTANCE and UNAVAILABLE, labels demo data and retains a user-resettable server identity binding.
• Local Server/Client role entrypoints in a wheel, canonical version operation from 2.3.0 to 2.4.0, source tests and installed-wheel qualification script.
• Owning backlog, architecture, roadmap and existing HTTP/distribution DAGs updated with candidate evidence. Full parent nodes stay PLANNED. WH-PEERS/full WH-Q and public WPK publication/readback/consumer gates remain open.

Exact local validation on candidate head 2492f378f919970cffbc2c3d37b2f6d932ba0e43

• bash scripts/validate.sh: PASS, including ten new runtime tests and six earlier lane-entry guards.
• Executable-line coverage per changed production module: __init__.py 7/7 (100%), cli.py 42/43 (97.67%), http.py 73/86 (84.88%), service.py 82/86 (95.35%). Each strictly exceeds 80.2%.
• Bandit static scan of workspace_control: no findings. git diff --check: PASS.
• Built pcvantol_workspace_control-2.4.0-py3-none-any.whl, SHA-256 45e173cfe5e8c483c06de9f868a25948b7464393dc56fae42739ad6aa8392ad7; installed non-editably in a fresh environment outside the checkout. Real Server API/CLI and headless Chrome Client interaction: PASS. Two isolated instances and normal SIGTERM/restart preserved identity/catalogue; no peer contacted.
• The first independent read-only review at prior head 1b81fa1 found a concurrent-init cleanup race. That P2 finding was fixed and regression-tested. Fresh independent read-only review at exact head 2492f37 verified the repair and found no remaining actionable issues. It was not a GitHub approval or reviewer grant.
• Exact-head protected Foundation validation and version-source jobs: SUCCESS. Non-blocking TDE observe job: SUCCESS; its artifact reports assessment PASS_WITH_WARNINGS, repository qualification QUALIFIED, process exits 1 and no coverage evidence. This is observation, not TDE qualification of the runtime.

The GitHub OAuth connection refused changes to .github/workflows/* without workflow scope. Those attempted workflow edits were removed before push; existing CI still runs the expanded scripts/validate.sh. The strict coverage/build/installed-browser proof above is local candidate evidence, not an added hosted CI gate.

This first result does not install a system service, use production credentials, contact Forge/EP, publish to PyPI, change a peer repository, run a Mission or require a reboot. Browser transport is loopback HTTP only; remote HTTPS and peer operations need later bounded qualification. The first installer DoD is unchanged.

Protected merge and installed main readback

PR #35 was squash-merged to protected main as 7d41bfad14a349ebed724773b5e1e4e213f3549b. Main has the same source tree as the reviewed head. Local bash scripts/validate.sh and version-source check passed on exact main. Main Workspace CI, canonical versioning and non-blocking TDE observe workflows completed SUCCESS. The main TDE evidence retained PASS_WITH_WARNINGS / repository QUALIFIED, process exits 1, and no TDE coverage evidence; the independent local per-file coverage above remains the applicable measured source proof.

A fresh wheel built from protected main was installed outside checkout and passed the same real API/CLI/headless-browser/two-instance/restart probe; its SHA-256 is 5984017f99b36566779c6d34a2848acfb100c456393e87442af006dab1b9898e. The wheel rebuild has different ZIP bytes from the reviewed-head wheel; no bitwise reproducibility claim is made. The local checkout is clean on main, and the merged feature branch and all own test processes/listeners were removed.