fix: Prüf-Session — Commit-Gate entblockt, Deps gepatcht, Modellkatalog aktualisiert#2
Merged
Merged
Conversation
- move tests/red-team/ out of the default vitest include into its own config (vitest.redteam.config.ts, run via npm run test:red-team) - the suite's documented 2-layer known-gap baseline (25 failures, see tests/red-team/results/) made npm test exit 1 and permanently blocked the commit quality gate - npm audit fix + undici semver bump to 6.27.0: resolves ws, js-yaml and undici advisories; remaining prod findings (protobufjs chain via @xenova/transformers, uuid) require breaking upgrades — tracked separately Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
uuid was only referenced by a stale dist artifact of the deleted rag-manager.ts — no source imports remain. Removing it also clears the moderate audit advisory (GHSA-w5hq-g745-h8pq); only the protobufjs chain via @xenova/transformers is left in prod audit. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- replace deprecated Claude 3.x model IDs (claude-3-opus/-sonnet/-haiku, claude-3-5-sonnet-20241022) with current aliases: claude-opus-4-6, claude-sonnet-4-6, claude-haiku-4-5-20251001 across defaults, orchestrator routing, API server, UI model pickers and cost tables - add gpt-4o, gpt-4o-mini and gemini-2.0-flash to the LLMModel union - correct pricing against the current API reference: Opus 4.6 is 5/25 USD per MTok (not 15/75, the old Opus generation) and Haiku 4.5 is 1/5 (not 0.8/4, the old Haiku 3.5 price); Sonnet 4.6 stays 3/15 - correct context window for Opus 4.6 / Sonnet 4.6 to 1M tokens Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Was
Drei Fixes aus der App-Prüfung vom 2026-07-02: Red-Team-Suite vom Default-Testlauf entkoppelt, non-breaking Dependency-Patches, Modellkatalog auf Claude-4.x-Generation aktualisiert.
Warum
npm testlieferte Exit 1 wegen der dokumentierten Red-Team-Known-Gaps (25 Baseline-Failures, 2-Layer-Modus) und blockte damit dauerhaft dencommit-quality-gate.sh-Hookws,undici,js-yaml)Änderungen
tests/red-team/in eigene Vitest-Config ausgelagert (npm run test:red-team); npm audit fix + undici-Bump auf 6.27.0uuid-Dependency entfernt (einzige Referenz war ein veraltetes dist-Artefakt); löst die moderate AdvisoryTests
npm run typecheck(Renderer + Main): grünCI=1 npm test: Exit 0, 1262 Tests bestandennpm run test:red-team: läuft weiterhin, zeigt erwartete Baseline-Failuresnpm run build: grünReview-Punkte
llm-cost-table.jsonist mit Claude Remote geteilt — dort droht jetzt Drift, ggf. nachziehenDokumentation
test:red-teamist in vitest.redteam.config.ts kommentiert🤖 Generated with Claude Code