Repository navigation
T3 follow-ups from PR #272's review: transport reach and writer-mode note #273
Description
Current scope
Two follow-ups from Unfret's review of PR #272's merged head:
v1:b780a83d…verdict-changing: a PR that changes only the T3 transport code (.claude/skills/verify-open-pstack/scripts/harness.tsT3 parts,scripts/doctor.tst3Binary) cannot namet3:dispatch, because classification reaches only Claude and Codexproject-skillfor those paths. Make those paths reacht3:dispatchas well, the wayrecipes.tsdoes.v1:39c9e028…: Eric decided on 2026-11-06 ("T3 for everything") that writer lanes in a T3 session keep going throughdelegate_taskand run in T3's runtime mode, which is full-access in his sessions, without the<provider> writer sandbox:opt-in. Document that in the## T3 hostsection ofplugins/pstack/skills/poteto-mode/references/provider-dispatch.md: the writer-sandbox setting governs only lanes that take the standalone route.
Acceptance:
- Item 1 has a table-test row that fails before and passes after.
- Item 2 is one or two sentences in the T3 host section; setup's warning text mentions it where it describes the setting.
- Live gate:
t3:dispatchwith the mandatory--self-test, nothing else.
Introduced by: #272
Activity
ericlitman commentedon Nov 6, 2026
PR: #274
Merge pull request #274 from ericlitman/fix/273-t3-followups
fix(verify): let T3 transport changes name t3:dispatch
docs(t3): say the writer-sandbox setting does not reach T3-routed wri…
ericlitman commentedon Nov 6, 2026
Batch 262 closeout for this issue (full record: https://github.com/mobilyze-llc/docs/blob/main/projects/open-pstack/records/2026-11-06-batch-262/closeout.md):
#273: T3 follow-ups from PR #272's review
Fix-forward PR #274, merged head f9bcda9 approved. T3 transport changes reach
t3:dispatch(reach test fails before, passes after); the T3 host section and setup's Codex question say the writer-sandbox setting does not reach T3-routed writers. Live:t3:dispatchplus self-test, 3 sessions.