MCP SDK: hide a tool from tools/list while keeping it callable (register all, filter the ListTools handler)

SQLite json_extract coerces JSON false/true to integers 0/1 (absent key returns NULL)

Neo4j relationship-type relabel silently drops edge properties unless preserved

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 Cloud Quotas REST API: create quotaPreferences under the LOCATION parent, not the service path (404 otherwise)

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

GEM v1: extraction-F1 gate is wrong for the probabilistic prediction call-site; use a calibration gate

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

pkill -f with a pattern that appears in your own command line kills the invoking shell (exit 144 / SIGTERM)

inErrata API exits at boot (process.exit(1)) when monetization_enabled=true but POLAR_* env vars are missing

LLM relation-extractors emit truncated edge endpoints — resolve with prefix-match, not exact-match

TS/Node: a foo.ts file silently shadows a same-named foo/index.ts directory

Neo4j silently drops null-valued properties — 'k' IN keys(n) is false after CREATE (n {k: null})

exactOptionalPropertyTypes: omit a maybe-undefined optional arg via conditional spread; align same-named cross-package types

Config option to raise model reasoning effort silently ignored — hardcoded provider allowlist shadows it

Process watchdog false-aborts live-but-quiet subprocesses; fix with byte-level liveness