Stop the search chain treating an empty answer as an answer

Classify a blocked Scholar page before waiting, not after

Make a failed Scholar scrape say what it actually got

Detect Google Scholar's block page, and unwrap Bing's redirect links

Move orchestration out of the route handlers into a service layer

Housekeeping: lint in CI, client lifecycle, and a truer breaker signal

Pin the MCP tool surface and give every tool an output schema

Test the search URLs, and delete the harness that hid them

Stop reporting upstream failures as "patent not found"

Bound the list inputs and the outbound fan-out they cause

Run the blocking DuckDuckGo call in a worker thread

Test the API-key middleware, and stop it 500ing on a non-ASCII key

Add a circuit breaker to the multi-backend search fallback (item from review)

Run the container as a non-root user (item from review: Dockerfile hardening)

Validate patent_id shape before hitting Google Patents/OPS

Deduplicate app.py's search endpoints (item 4 from review)

Split fetch from parse in scrap.py and serp.py (item 2 from review)

Fix bare except and misattributed logger (item 3 from review)

Fix hanging Playwright tests: make the browser fixture function-scoped

Add pytest harness with characterization tests for scrap/ops/serp/utils