Give your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
Python agent-infrastructureai-agentai-searchautomationbilibiliclaude-codeclicursorfree-apillm-toolsmcppythonreddit-scrapertwitter-scraperweb-scraperxiaohongshuyoutube-transcript
docs: record installed wheel and shared browser acceptance
LohseZSaid598569de34b
diff --git a/docs/validation/shared-browser-20261020.md b/docs/validation/shared-browser-20261020.mdindex 0792b7d6..b38c5fd1 100644--- a/docs/validation/shared-browser-20261020.md+++ b/docs/validation/shared-browser-20261020.md@@ -45,8 +45,10 @@ Official 1.8.8 failed the Xiaohongshu default-filter path with `UPSTREAM_BROKEN`## Validation- Core tests: **384 passed** on local macOS.+- The curated source checkout also passed all 384 tests, with local API/research files excluded.- Ruff: passed. Mypy: passed, 46 source files.-- Wheel gate: requires Skill reference files and Xiaoyuzhou; rejects removed planner/MCP/extractor/patch modules.+- Wheel gate passed: 57 entries, about 148 KiB; Skill references and Xiaoyuzhou present, removed modules and browser binaries absent. Clean-venv smoke test confirmed 14 channels and rejection of the four retired command entries.+- Installed the candidate wheel locally, ran setup successfully (5/5 defaults), refreshed the installed Skill and confirmed no retired Reach MCP configuration/process. An actual Xiaohongshu search through the installed wheel returned one item in 7.48 seconds.- Packaging uses reviewed source files only. Local API prototypes, provider research, account data, browser profiles and downloaded binaries are excluded.## User acceptance