Remove the premature visibility classifier and let one proof agent configure and exercise the disposable Telegram gateway. Align mock response timing with the 15-minute lane budget while preserving credential isolation through the alias-token proxy.
Move Mantis Telegram Desktop proof from the remote AWS/Crabbox lane to a recorder-driven local Docker desktop. Keep proof scenarios agent-authored, cache trusted build outputs, and publish exact visible Telegram evidence without writing the QA bot token to artifacts.
Co-authored-by: Ayaan Zaidi <hi@obviy.us>
* refactor(agents): reduce tool failure warnings to two rules
* test(agents): remove obsolete tool recovery receipt proof
* refactor(agents): drop unused meta param from buildToolMutationState
* test(agents): remove stale tool warning assertions
* test: export runtime source snapshot from closed runtime-snapshot mocks
Heal the main breakage introduced by 9441e3fe6e / #126531, which added a runtime source-snapshot read to provider model route resolution. Closed Vitest factories now return null for that source snapshot, preserving their pre-projection behavior.
* test(gateway): make compaction read-error faults order-immune
Generation-2 CI failure in run 32342180898, job 96343444772 showed that the mock factory initialized while shared gateway-server importers remained bound to the real transcript reader.
The dedicated isolated project fixes normal shards. Complete its ownership by adding it to the root project matrix and excluding the test from the non-isolated OPENCLAW_GATEWAY_PROJECT_SHARDS=0 fallback.
Preserve attempt-local recovery state so successful mutation retries emit a redacted terminal receipt without stale failure warnings.
Co-authored-by: Ayaan Zaidi <hi@obviy.us>
The mock streamed pre-tool assistant text on the Responses path without `phase: "commentary"`, the field the API sets and the transport reads, so the draft-proof scenario ran green while never exercising the preamble lane on that transport. Proven live on Telegram: the preamble now renders as the status headline, and as a commentary line when progress.commentary is on.
A progress draft only exists while the agent is working, and OpenClaw's exec
returns instantly for a mock tool call, so the whole turn finished in ~86ms —
inside the 1.5s draft gate. No draft was created, which is correct behavior but
proves nothing about rendering.
The draft-proof scenario now holds its final answer (MOCK_DRAFTPROOF_FINAL_DELAY_MS,
default 6s) so the turn straddles the gate. That is what made the live Discord
recording possible: draft created, edited with the status headline above the
tool line, replaced by the final answer, then deleted.
Three gates my own changes broke:
- docs/channels/telegram.md needed a blank line before </Note>; I formatted the
TypeScript but never ran the docs formatter over the pages I edited.
- The mock harness matched a literal with a regex, which oxlint's prefer-includes
rejects.
- Widening resolveChannelPreviewStreamMode's defaultMode to StreamingMode and
adding rendersRollingLinesNatively changes the exported plugin-SDK surface, so
the API baseline hash had to be regenerated.
Proving headline-plus-tool-line composition needs assistant text emitted
*before* a tool call: that leading text is what gets tagged as commentary and
rendered as the draft's status headline. The mock could only echo a single token,
so the draft path was unreachable without a live model.
Trigger `OPENCLAW_E2E_DRAFTPROOF` now serves two turns — preamble plus a shell
call, then the final answer — on both /v1/responses and /v1/chat/completions.
Only the completions transport tags commentary today, which is why the
completions path matters here.
Used to record the draft live on real Telegram: one message carrying the
headline above the tool line, the final answer, then the draft edited to its
receipt.
Ayaan Zaidi