Starstruk Logo MarkStarstruk

v0.1.79

Skyvern-AI/skyvernv0.1.79Apr 30, 2025by wintonzheng

AI Summary

A substantial update introducing the new 'Block Library' terminology, expanding AI model support to include Bedrock and Vertex AI, improving CUA safety and action handling, and adding self-hosted proxy capabilities.

Key Highlights

  • Renamed Node Library to Block Library
  • Added support for Bedrock Claude 3.7 Sonnet and Vertex AI
  • Enhanced CUA with safety checks and action handlers (SOLVE_CAPTCHA, CLICK, INPUT_TEXT)
  • Added self-hosted proxy support

New Features

  • Block Library
  • New AI models (Bedrock, Vertex)
  • CUA safety & actions
  • Proxy support

Full Release Notes

## What's Changed
* Node Library -> Block Library by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2183
* safety checks for CUA by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2188
* fix CUA pending checks schema by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2189
* Block Library naming fix by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2190
* Feature/self hosted proxy support by @piyushmittal20 in https://github.com/Skyvern-AI/skyvern/pull/2115
* acknowledged_safety_checks by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2192
* disable async operations for CUA agent by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2194
* remove terminate in cua fallback by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2195
* terminate action when CUA verification code is not received by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2196
* do not take fullpage screenshot post CUA action by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2197
* upgrade litellm to 1.67.0 by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2198
* support workflow_run_id in totp code by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2199
* Feature: Configuring OpenAI-Compatible (LiteLLM) Models by @TheDropZone in https://github.com/Skyvern-AI/skyvern/pull/2166
* add workflow_run_id in totp endpoint doc by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2200
* fix custom selection bug by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2201
* update latest openai model configs for azure & openai by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2202
* bedrock claude 3.7 sonnet by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2203
* support shadow dom mutation observer by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2206
* update need verification prompt by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2207
* fix prompt typo by @LawyZheng in https://github.com/Skyvern-AI/skyvern/pull/2208
* log skyvern's response to CUA questions by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2209
* add more scrape page methods by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2211
* handle null tagname by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2212
* fix task v2 block totp issue by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2214
* optimize and speed up custom selection by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2215
* fix phone number format issue by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2217
* Disable connection pooler if env specifies it by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2222
* debug log: Using general model configuration for unknown LLM key by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2223
* bump v0.1.78 by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2224
* add LLMCaller that supports message history by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2204
* optimize max steps failure reason for task v2 by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2225
* Fix Skyvern.node.ts for N8N Community Node by @dd36 in https://github.com/Skyvern-AI/skyvern/pull/2216
* bump up n8n version to 0.0.3 by @LawyZheng in https://github.com/Skyvern-AI/skyvern/pull/2226
* add current sha into log by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2227
* fix interactable parsing by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2228
* anthropic CUA by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2231
* increase download max wait time and change logs by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2232
* do not close the tab when download is not triggered by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2234
* improve custom selection by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2236
* use anthropic beta client by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2235
* bugfix by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2237
* Anthropic CUA fix by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2238
* fix anthropic messages image input pattern by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2239
* fix anthropic llm caller routing by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2241
* enable thinking in anthropic CUA by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2240
* fix tool_call_input parsing with anthropic sdk by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2242
* add vertex ai by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2243
* add google dependency by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2244
* Add organization_name and workflow_permanent_id to skyvern context, pass workflow_permanet_id when deciding which identifier to use with which llm by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2245
* fix anthropic tool result format by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2246
* fix bug by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2249
* support only selectable input element by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2250
* Show continue_on_failure in loop node by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2251
* Update docker compose with Bitwarden by @suchintan in https://github.com/Skyvern-AI/skyvern/pull/2253
* bugfix by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2254
* resize screenshots to make it compatible with anthropic CUA by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2255
* CUA action parser - generate_cua_fallback_actions helper by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2256
* remove mintlify doc by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2097
* support more anthropic actions by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2257
* action handlers back - SOLVE_CAPTCHA, CLICK, INPUT_TEXT by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2258
* bump version to v0.1.79 by @wintonzheng in https://github.com/Skyvern-AI/skyvern/pull/2259

## New Contributors
* @TheDropZone made their first contribution in https://github.com/Skyvern-AI/skyvern/pull/2166
* @dd36 made their first contribution in https://github.com/Skyvern-AI/skyvern/pull/2216

**Full Changelog**: https://github.com/Skyvern-AI/skyvern/compare/v0.1.77...v0.1.79