[Cuarroc]

Dieses Paket fügt nur eine ungenutzte, reine Datenfunktion hinzu: Sie trennt eingestellte Modell-Hinweise von echten, mit Quelle und Zeit belegten Beobachtungen. Unbekannte Werte für Aufwand und Kosten bleiben "unbekannt" und werden nie zu 0 oder "gratis"; nur ein ausdrücklich beobachtetes 0 bleibt 0. Es gibt keine Oberfläche, keinen Aufrufer, keine Modellwahl und keine Quoten- oder Schlüsseländerung, die Modellauswahl-Oberfläche (C08UI) bleibt hinter A1+R2 gesperrt.

Report

Package: V36-C08-DATA (parent V36-C08), Tier B, operator UMSETZEN, DRAFT. Base 82a79d25, head 0d373808. Parent M0-P2 (PR #2015, merge b1645626) is an ancestor of the base (checked, exit 0). Authority: Root/Stabs controller admission of the pure unused DATA leaf; no R2/Satz3/paid/key/routing grant.

What changed (RAW 178 lines vs base, cap 180: 121 src/lib/providers.ts + 57 src/lib/providers.test.ts, 0 deletions)

Additive only; no existing function, type, hook, export or test body was touched, src/types.ts untouched (read-only), no live caller added.

New exports in src/lib/providers.ts:

• ModelFact<T> = {state:"known",value} | {state:"unknown"}
• ModelFactsInput {providerId, harness, configured?{model?,effort?}, observed?{providerId,harness,sourceId,model,observedAt (unix s),effort?,cost?{amount,unit}} | null}
• ModelFactsClock {nowSeconds, staleAfterSeconds} (caller-supplied; no Date.now, no built-in TTL)
• ModelFacts {configuredHint{model|null,effort|null}, observed: {state:"unknown",reason:"missing"|"identity-mismatch"|"invalid"|"future"} | {state:"observed",freshness:"current"|"stale",sourceId,model,observedAt,effort:ModelFact<string>,cost:ModelFact<{amount,unit}>}}
• projectModelFacts(input, clock): ModelFacts: pure; reads own data properties only (no getters, no spread, no logging, inherited members ignored). Represents evidence, does not authenticate the caller.

Literal oracle table (fixed in the tests, not computed by production code):

Case	Result
configured model-configured/high, no observation	hint kept, observed unknown/missing
observed model-observed, source session-1, t=900, now=1000, window 300, effort/cost null	observed/current, source+time kept, effort and cost unknown
observed with effort:"low", cost {0,"usd"}	effort known low, cost known {0,"usd"} (explicit 0 preserved)
other providerId / other harness	unknown/identity-mismatch
t=1001 (> now)	unknown/future
t=0, NaN, blank sourceId, empty model, NaN clock	unknown/invalid
t=700 / t=699 (window 300)	current / stale (model still reported, non-current)
cost amount -1, missing unit, string amount, missing amount	cost unknown
getters on effort/cost/configured, unrelated field secret	getters never run (0 reads), field not in output

Evidence (all foreground under the assigned guard and warm slot, CARGO_BUILD_JOBS=1)

• Stub checkpoint 7edad0ff (same public types, stub projection): npm run typecheck exit 0; npm run test:unit -- src/lib/providers.test.ts -t "keeps configured model hints distinct from observed evidence and unknown cost" exit 1, 1 failed / 23 skipped, AssertionError on configuredHint (assertion red, count 1, not a missing-import red). Committed through the normal hooks.
• Head 0d373808: same selector exit 0 (1 passed / 23 skipped); whole src/lib/providers.test.ts exit 0 (24 passed); npm run typecheck exit 0.
• Neighbours importing providers (ShellAccessibility, UsageView, UsageView.claims, ErsteinrichtungScreen tests): exit 0, 25 passed.
• bash scripts/ci/red-first.sh (canonical, unchanged, own internal targets): exit 0, all 3 Test-First entries red/missing at base and green at head.
• bash scripts/ci/gates.sh lane prepush in this worktree: exit 0 (313 s; typecheck, lint, fe-test, hq-test, denkraum, clippy, rust-suite green).
• Normal commit hooks (secret scan, red-first, commit-msg) and push hook enabled, no --no-verify; git push -u exit 0, git ls-remote shows 0d373808bde02bcc69498effbb15a82a3e570434.
• Raw diff is additive: git diff --numstat 82a79d25 gives 57/0 and 121/0.

Test names: the second case was named without a comma so the Test-First trailer is not read as a list. No old assertion was changed.

Architecture drift

No finding; no seam file touched, no allowlist change.

NICHT ABGEDECKT

• Windows lane (#[cfg(windows)] tests, ConPTY/DPAPI, Windows clippy arms): not covered on Linux; it runs in the merge queue.
• Prepush is the fast loop: browser smoke, frontend build and workflow gates run only in the linux lane.
• No live model, usage, cost, R2 or selection proof; no visual claim (data-only, no screenshot, no UI critic).
• No caller consumes the projection yet; real observation sources, their authenticity and any TTL choice belong to C08UI/R2.
• The model Co-Authored-By is Claude Sonnet 5.5 (native header); effort was not observed.
• External service state is checked by no gate.

Review

Review: ausstehend (Stufe B, zwei verschiedene Nicht-Autor-Familien, vom Orchestrator/Root eingereiht). No author review launched, no Ready/queue/merge. Stabs updates PLAN after merge; C08UI stays open and gated.

🤖 Generated with Claude Code

Root review record (2026-10-30 00:25Z, head 0d37380)

Source	Family	Verdict
Codex gpt-6.1-sol/high (server review feeder job srv-review-pr2041-codex-0104)	OpenAI	APPROVE WITH CONDITIONS
Kimi Code CLI (Root, PC)	Moonshot	APPROVE WITH CONDITIONS (no high/medium)

ID	Sev	Finding	Disposition
R2041-P1 (Codex)	med	PLAN row stale (docs/PLAN.md:406)	waived: Stabschef PFLEGEN after merge (standing practice)
R2041-A1 (Kimi)	low	nonBlank keeps untrimmed value; asymmetric trim in identity compare (providers.ts:506-510)	follow-up for C08 UI wiring (no caller today)
R2041-A2 (Kimi)	low	docstring "no clock is read" wording (providers.ts:500-501)	follow-up
R2041-B1 (Kimi)	low	as never cast in test 2	accepted as intentional runtime-validation test

Tier B complete with two non-author families (author Anthropic) → ready + review-ok bound to 0d37380.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-30T00:18:49.777014Z	—	Draft marked ready

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[mergify]

✅ Merged into main as 227e10d at 00:33 UTC

9 min in the queue, 8 min of it CI.

So far

• ✅ Entered the queue Fri 30 Oct, 00:24 · rule default · triggered by merge protections
• ✅ Passed · draft merge queue: merged #2041: feat(providers): V36-C08-DATA inactive model/evidence facts projection #2042
• ✅ Merged 00:33 · 227e10d · merge

Merge conditions: 9 of 9 met

This pull request

	Condition	Required by
✅	github-review-approved	GitHub repository ruleset rule main-pr-required-checks
✅	label != do-not-merge	—

The batch, tested on draft #2042

	Condition	Required by
✅	check-success = gates (linux)	—
✅	check-success = gates (windows)	—
✅	check-success = red-first	—
✅	any of: check-success = @github-actions/gates (linux), check-neutral = @github-actions/gates (linux), check-skipped = @github-actions/gates (linux)	GitHub repository ruleset rule main-pr-required-checks
✅	any of: check-success = @github-actions/gates (windows), check-neutral = @github-actions/gates (windows), check-skipped = @github-actions/gates (windows)	GitHub repository ruleset rule main-pr-required-checks
✅	any of: check-success = @github-actions/red-first, check-neutral = @github-actions/red-first, check-skipped = @github-actions/red-first	GitHub repository ruleset rule main-pr-required-checks
✅	any of: check-neutral = @mergify/Mergify Merge Protections, check-skipped = @mergify/Mergify Merge Protections, check-success = @mergify/Mergify Merge Protections	GitHub repository ruleset rule main-pr-required-checks

_{Times in UTC · dashboard · how the merge queue works}