test(google): guard Argon thought separation across client protocols

devlog/_fin/261015_argon_thought_boundary/000_plan.md

@@ -0,0 +1,97 @@
+# Gemini Argon output separation
+
+A report of reasoning-looking assistant output prompted comparison of the Google response
+classifier with CLIProxyAPI and cc-switch. The inspected classifier already separates marked
+thoughts. This unit verifies that separation through every client protocol using synthetic
+markers, and documents the boundary without claiming the original observation is reproduced.
+
+## Loop specification
+
+- Archetype: satisfy-spec; one C3 verification/documentation work phase.
+- Trigger/goal: investigate Argon output separation and deliver an evidence-backed PR to dev.
+- In scope: synthetic Argon integration tests, their layout registration, and current Google docs.
+- Non-goals: heuristic text censorship, runtime changes without a demonstrated defect, private
+  session transcripts, user configuration changes, deployment, other worktrees, native PR stacks.
+- Verifier: new protocol test plus adjacent Google/direct-encoder tests; typecheck, test layout,
+  file-size, privacy and structure gates; docs build; required CI at the PR head.
+- Stop: coverage/docs merged with passing relevant checks, or a concrete blocker recorded.
+- Memory artifact: this unit and local `.tmp/argon-protocol/` evidence; no memory-store writes.
+- Outcomes: DONE for the verification PR; runtime NOOP if marked thoughts remain separated.
+  Do not claim the reported live symptom fixed merely because synthetic tests pass.
+- Escalation: a demonstrated defect changes the plan before patching; unavailable live evidence
+  remains an explicit limitation. Existing user authority includes push/PR/merge, no release.
+- Bounds: none stated; host limits apply. Tools: local Bun, read-only public Aside/GitHub research,
+  repository-scoped GitHub publication. No production credentials are needed for tests.
+
+## Evidence and hypotheses
+
+- H1: the adapter converts marked thoughts to answer text. Falsifier: mixed marked/unmarked
+  synthetic parts preserve separate channels in streaming and buffered responses.
+- H2: a client converter collapses channels. Falsifier: actual Responses, Chat and Messages
+  handlers keep thought markers out of ordinary answer fields with direct routing on and off.
+- H3: upstream ordinary text or client rendering explains the observation. Static tests cannot
+  establish this; never reclassify prose by its wording or infer a root cause from test success.
+
+Source anchors: `src/adapters/google.ts:716`, `:989`, `:1378`, `:1654`;
+`src/protocols/encoders/chat.ts:133`; `src/protocols/encoders/messages.ts:135`.
+CLIProxyAPI `5fc580b1d72d87a7bcfe5cfb6f479c1aea5625b3` maps `thought` to
+`reasoning_content`; cc-switch `76f99d41a66e24a5ed2e8982bc494399f306bfe6` filters
+marked thoughts from visible text. Public Aside reads succeeded with complete=true,
+truncated=false, contentVerified=true for Google thinking docs and CLIProxyAPI #6607.
+Those docs distinguish thought signatures from ordinary text visibility; #6607 is a buffering
+decision, not evidence of text leakage. Argon upstream behavior is not independently certified.
+
+Baseline: `bun test tests/adapters/google/google-hardening.test.ts
+tests/adapters/google/google-adapter.test.ts tests/adapters/google/gemini-4-argon-antigravity.test.ts
+tests/responses/responses-show-thinking-summary.test.ts`: exit 0, 158 pass at 7640a942b3.
+Work then refreshed to dev 0b13396af4; adjacent checks will run on the final patch.
+Typecheck's `include: ["src"]` does not validate new tests; Bun execution does.
+
+## Consultation
+
+Architect proposal and reflection are retained in the local coordinator evidence.
+- ARGON-SEP-01/02 accepted: keep one classifier and existing output channels.
+- ARGON-SEP-03 accepted: provenance is not proof of observed upstream behavior.
+- ARGON-SEP-04 amended: add handler-level coverage and current contract clarification.
+  Preserve historical ADR-0055 per `structure/AGENTS.md`; do not rewrite the old decision.
+- Same-architect reflection: ALIGNED against `010_verification.md`; require assertions on both
+  deltas and final accumulated payloads, and correct tool signature attachment where exposed.
+  Independent plan audit: PASS. Refreshed-base adjacent direct/encoder baseline: 131 pass, 0 fail.
+
+No new runtime fields or enums. No new runtime enforcement: tests are regression checks only;
+they can be bypassed by skipping tests, and cannot certify external client rendering.
+
+## Verification result
+
+The investigation's runtime outcome is NOOP: the tested marked thoughts stay out of ordinary
+answer fields. New coverage adds 75 cases. Deliberately changing the classifier to emit marked
+thoughts as text fails all 75; byte-for-byte restoration returns all 75 to green. No classifier
+mutation is part of the patch. The live observation's cause remains unresolved.
+
+Final focused validation on the repository-pinned Bun 1.4.3: 364 pass / 0 fail across the new
+matrix and eight adjacent Google/direct-protocol files; layout/tooling/size: 36 pass / 0 fail.
+`bun run test:changed` selects the new file only and passes all 75 cases. `bun run typecheck`,
+`bun run structure:check`, and `bun run privacy:scan` pass. `cd docs-site && bun run build`
+builds 577 pages and checks 81,202 links. Raw logs stay in local `.tmp/argon-protocol/`.
+An initial layout run under the shell's Bun 1.4.0 failed the pinned-runner check; using the
+already-installed Bun 1.4.3 resolved it without source changes.
+
+Initial fixture expectations incorrectly treated a disabled provider display default as a
+client opt-out and omitted the Messages thinking opt-in. The current tests explicitly request
+the protocol's display state, matching `core-normalize.ts` and `claude/inbound.ts`; they still
+force upstream marked thoughts when hidden to test output suppression, not just request flags.
+
+The full 2,330-file suite is not run locally: this tests/docs-only change has bounded integration
+coverage and shares the host with concurrent coding workloads. Broader coverage remains with
+the PR's required CI. Neither this exception nor synthetic tests certify external harness UI.
+
+Fresh independent final review: PASS, no actionable defects. The fresh reader identified the
+result as synthetic regression coverage, with the live symptom explicitly unverified. No prose
+change was requested. The investigation is closed with that limitation; PR integration is
+tracked by the coordinator. Runtime source is unchanged.
+
+Hosted review follow-up: restore all fixture globals in finally if spend-lease release throws;
+add independent client-only opt-outs with provider opt-in still true for all three protocols.
+A Chat opt-out mutation fails 18 cases; a Messages display-omitted mutation fails 6. All
+production files were restored byte-for-byte. These probes confirm the added cases detect
+the independent client preference, rather than passing only because the provider default is off.

devlog/_fin/261015_argon_thought_boundary/010_verification.md

@@ -0,0 +1,33 @@
+# One phase: verify the Argon output boundary
+
+Depends on current Google adapter and existing client handlers; no runtime design changes.
+
+1. NEW `tests/responses/protocol-argon-thought-boundary.test.ts`: isolated homes and synthetic
+   credentials, mock Google transport only, invoke real Responses/Chat/Messages handlers.
+   Generate `thought: true` marker parts, ordinary text (absent and false markers), signature-only
+   parts, an unmarked signed text part, and an optional tool call with synthetic signature.
+   Exercise buffered and SSE, coalesced and split SSE frames, and direct/bridge Chat/Messages.
+   Assert exact ordinary answer text in both deltas and accumulated final payloads,
+   thought marker only in reasoning/thinking fields, no
+   signature-only text, preserved tool name/args and call-bound signature where the client wire
+   exposes it, successful terminal. For independent client-only opt-outs (Responses/Chat summary none and Messages display omitted),
+   force upstream thought parts anyway and require suppression while answer and tool survive.
+   Inspect protocol trace so the direct/bridge cases prove the intended lane executed.
+2. MODIFY `scripts/test-layout/layout.json` and `tests/fixtures/test-layout-expected.json`:
+   register the new test in responses, using existing shard conventions.
+3. MODIFY `structure/providers/google.md`: describe signature-independent text classification
+   and reference the new test; preserve the historical decision record.
+4. MODIFY `docs-site/src/content/docs/guides/providers.md`: clarify Gemini thought summaries,
+   client field separation, and that unmarked ordinary text is not censored by wording.
+5. Run new tests and adjacent Google/direct-lane/encoder tests. Validate the new oracle by a
+   temporary local classifier mutation (marked thoughts become text), require failure, restore
+   bytes, then require green. No mutation may be committed.
+6. Run typecheck, test-layout/tooling, file-size, privacy, structure and docs build. Full suite
+   default applies; if concurrent workload makes it disproportionate for tests/docs only, record
+   exact focused scope and leave remaining platform/integration coverage to required CI.
+7. Fresh independent review, fix valid findings, create ordinary PR to dev using template.
+   Inspect exact-head required CI and maintainer-integration policy before the authorized merge.
+
+Acceptance is protocol correctness for the synthetic fixtures and prevention of regression.
+The original live symptom and other harnesses' rendering remain unverified; the PR and final
+report must say so. A newly demonstrated runtime defect requires a plan amendment and audit.

docs-site/src/content/docs/guides/providers.md

@@ -157,6 +157,14 @@ eligible Enterprise accounts. Cloud Code Assist serves Argon only to the Antigra
 OpenCodex sends Argon requests and model discovery with the CLI User-Agent
 (`antigravity/cli/1.3.4`) and keeps the IDE User-Agent for every other model.
 
+For Gemini on Antigravity, including Argon, provider-marked thought summaries stay separate
+from the answer: Responses uses reasoning summary items, Chat Completions uses
+`reasoning_content`, and Messages uses `thinking` blocks. Clients decide how to display those
+fields. A Responses client can request `reasoning.summary: "none"` to suppress summaries;
+this does not turn off the model's internal reasoning. A thought signature is continuation
+metadata and can accompany ordinary text, so its presence alone does not hide that text.
+OpenCodex preserves unmarked answer text even when its wording sounds like planning or analysis.
+
 The built-in `google-antigravity` model list is Gemini 3.8 Flash (`gemini-3.8-flash`), Gemini 4 Argon
 (`gemini-4-argon`), Claude Opus 5.5 (`claude-opus-5-5`), and Claude Sonnet 5.5
 (`claude-sonnet-5-5`), each with low, medium, and high efforts. What you actually see is your

scripts/test-layout/layout.json

@@ -2,7 +2,7 @@
   "version": 1,
   "root": "tests",
   "explicitShards": ["layout-clients.json"],
-  "explicit": { "service-log.test.ts": "service", "messages-native-empty-inline.test.ts": "claude-integration",
+  "explicit": { "protocol-argon-thought-boundary.test.ts": "responses", "service-log.test.ts": "service", "messages-native-empty-inline.test.ts": "claude-integration",
     "owner-registry-isolation.test.ts": "config",
     "codex-history-pending-wal.test.ts": "codex-integration", "responses-anthropic-billing-send-budget.test.ts": "responses", "anthropic-billing-ownership.test.ts": "adapters/anthropic", "server-anthropic-billing.test.ts": "server", "server-anthropic-billing-boundaries.test.ts": "server", "router-anthropic-billing-failover.test.ts": "routing", "anthropic-billing-refusal.test.ts": "adapters/anthropic", "messages-request-id-headers.test.ts": "claude-integration", "bounded-file-read.test.ts": "lib", "responses-undeclared-tool-guard-inbound.test.ts": "responses", "explicit-empty-tool-catalog.test.ts": "responses", "explicit-empty-tool-catalog-history.test.ts": "responses",
     "messages-request-id-endpoint.test.ts": "claude-integration",

structure/providers/google.md

@@ -11,6 +11,14 @@ explicit wire summary none win. Eligible CCA Gemini requests use `includeThought
 when provider opt-in and per-request display both allow it. Thought signatures remain attached
 to their tool calls independently; they never become Anthropic thinking signatures.
 
+The visibility marker is the boolean `thought: true`, not the presence of a `thoughtSignature`
+or the wording of the text. An unmarked signed text part remains ordinary assistant output;
+a signature-only part produces no text. Responses keeps summaries in reasoning items, Chat in
+`reasoning_content`, and Messages in thinking blocks. The handler-level synthetic matrix in
+`tests/responses/protocol-argon-thought-boundary.test.ts` checks Argon across buffered/SSE
+clients, coalesced/split upstream frames, direct/bridge lanes, explicit display suppression and
+tool calls. It verifies protocol separation, not live model behavior or another client's UI.
+
 > Decision record: [ADR-0055](../decisions/ADR-0055-google-thought-text-visibility-boundary.md)
 
 ## Google response-part field boundary

tests/fixtures/test-layout-expected.json

@@ -1,4 +1,5 @@
 {
+  "protocol-argon-thought-boundary.test.ts": "responses",
   "messages-native-empty-inline.test.ts": "claude-integration",
   "explicit-empty-tool-catalog.test.ts": "responses",
   "explicit-empty-tool-catalog-history.test.ts": "responses",