Change8

autogpt-platform-beta-v0.6.54

📦 autogptView on GitHub →
27 features🐛 31 fixes🔧 11 symbols

Summary

Release v0.6.54 introduces significant new features focused on CoPilot enhancements, agent execution modes, and credential management, alongside numerous bug fixes and maintenance updates including Codecov integration.

Migration Steps

  1. Rename placeholder_values to options on AgentDropdownInputBlock (if applicable to your code interacting with this block).

✨ New Features

  • #12515 - Replace suggestion pills with themed prompt categories
  • #12571 - Add copy button to user prompt messages in CoPilot
  • #12526 - Spend credits to reset CoPilot daily rate limit
  • #12577 - Add shared admin user search + rate-limit modal on spending page
  • #12583 - Live timer stats with persisted duration in CoPilot
  • #12512 - Add extended thinking execution mode to OrchestratorBlock
  • #12537 - Generic managed credential system with AgentMail auto-provisioning
  • #12582 - Add session-level dry_run flag to autopilot sessions
  • #12647 - Add generic ask_question copilot tool
  • #12650 - Set git committer identity from user's GitHub profile in E2B
  • #12569 - Add SQL query block with multi-database support for CoPilot analytics
  • #11797 - Update classic AutoGPT for improved day-to-day usability
  • #12655 - Set up Codecov coverage reporting across platform and classic
  • #12640 - Redesign onboarding wizard with Autopilot-first flow
  • #12622 - Add include_graph option to find_library_agent
  • #12581 - Add rate-limit tiering system for CoPilot
  • #12649 - Add gitleaks secret scanning to pre-commit hooks
  • #12578 - Add create → dry-run → fix loop to agent generation
  • #12672 - Add all 12 Z.ai GLM models via OpenRouter
  • #12665 - Add Playwright E2E coverage reporting to Codecov
  • #12623 - Add Fast/Thinking mode toggle with full tool parity in CoPilot
  • #12667 - Set up React integration testing with Vitest + RTL + MSW
  • #12629 - Add copilot artifact preview panel
  • #12328 - Load copilot messages from newest first with cursor-based pagination
  • #12691 - Add /orchestrate skill — Claude Code agent fleet supervisor with spare worktree lifecycle
  • #12696 - Cost tracking for system credentials
  • #12591 - Add /open-pr and /setup-repo skills

🐛 Bug Fixes

  • #12548 - Prevent graph execution stuck + steer SDK away from bash_exec
  • #12560 - Fix prod Sentry errors and reduce on-call alert noise
  • #12514 - Add sink input validation to AgentValidator
  • #12551 - Move placeholder_values exclusively to AgentDropdownInputBlock
  • #12561 - Upgrade Docker Node.js from v21 (EOL) to v22 LTS
  • #12565 - Fix prod Sentry errors and reduce on-call alert noise
  • #12574 - Preserve thinking blocks during transcript compaction
  • #12593 - Raise AIConditionBlock errors instead of swallowing them
  • #12579 - Resolve host-scoped credentials for authenticated web requests
  • #12555 - Disambiguate duplicate tool names in OrchestratorBlock
  • #12546 - Validate email recipients in Gmail blocks before API call
  • #12540 - Add empty choices guard in extract_openai_tool_calls()
  • #12545 - Validate non-empty input in AIConversationBlock before LLM call
  • #12604 - Add tool call circuit breakers and intermediate persistence in CoPilot
  • #12580 - Fix dry-run simulation showing INCOMPLETE/error status
  • #12606 - Remove LaunchDarkly local storage bootstrapping
  • #12632 - Prevent duplicate block execution from pre-launch arg mismatch
  • #12625 - Prompt-too-long retry, compaction churn, model-aware compression, and truncated tool call recovery
  • #12628 - Fix copilot credential loading across event loops
  • #12642 - Detect prompt-too-long in AssistantMessage content and ResultMessage success subtype
  • #12646 - Fix tool output file reading between E2B and host
  • #12635 - Recognize Agent tool name and route CLI state into workspace
  • #12575 - Propagate dry-run mode to special blocks with LLM-powered simulation
  • #12656 - Use OpenRouter client for simulator to support non-OpenAI models
  • #12660 - Prevent duplicate side effects from double-submit and stale-cache race
  • #12688 - Skip AI blocks without model property in fix_ai_model_parameter
  • #12709 - OrchestratorBlock dry-run credentials + Responses API status field
  • #12708 - Suppress Sentry noise from expected 401s in OnboardingProvider
  • #12707 - Make User RPC model forward-compatible during rolling deploys
  • #12713 - Wrap PlatformCostLog metadata in SafeJson to fix silent DataError
  • #12714 - Convert AttributeError to ValueError in _generate_schema

Affected Symbols