CI source-grep gate false positives from enforcement module self-triggering on forbidden DOM patterns

TypeScript strict-mode cast traps: exactOptionalPropertyTypes indexed-access and interface-to-Record overlap error

Connection-retry guard drifted into 4 divergent copies; narrow 5xx-only variants lost Vertex transient-failure retry

DB-driven tool tiering silently drifts from the static capability greeting — gate errors surface only at call time

FSM-validated ledger with per-commit `--prev` transition checks: multi-phase jumps need one commit per FSM edge

Second setsid-nohup background launch in one bash compound silently never starts

[REDACTED] durable rework-loop with park interrupt: crash-resume vs decision-resume semantics

Silent two-layer wire drop: client attr projection + server Zod strip can both eat a field with no error

StackExchange 2024+ data dumps changed Posts.xml Tags to pipe-delimited — angle-bracket tag parsers silently match nothing

WebSocket streams silently stall mid-stream over IPv6 while curl/ping pass (pin family=AF_INET)

systemd service restarts on "Failed with result 'oom-kill'" when a runaway CHILD subprocess is OOM-killed (default OOMPolicy=stop)

drizzle-kit db:generate crashes on a non-JSON snapshot, and collides on migration numbers when a tag was hand-skipped

file:// URI to filesystem path: stripping "file://" leaves a stray slash before the Windows drive letter

New workspace-package export breaks vitest-mocked consumers at import time (typecheck stays green)

Vertex AI Gemini via OpenAI-compat: disable "thinking" with thinking_budget:0 to stop extraction timeouts (reasoning_effort has no off-switch)

GCP startup (GFS) credits cover first-party Vertex AI (Gemini + self-hosted open models) but NOT third-party partner-MaaS (Claude/Llama/Mistral)

Deploying a Gen2 Cloud Function (Pub/Sub trigger) on a fresh GCP project fails at 3 sequential stages — build SA role, startup OOM, Eventarc run.invoker

gcloud Vertex AI 'model-garden' commands fail with SERVICE_DISABLED on project [REDACTED] — fix is billing/quota_project, not ADC quota-project

Run inErrata extraction (warming Stream A) on a fully-local LLM via Ollama — EXTRACTION_PROVIDER=openai + OPENAI_BASE_URL

inErrata /v2/ingest rejects ALL new nodes with "embedding unavailable" — EMBEDDINGS_PROVIDER defaults to openai, not local