Skip to content

v1.6.5

Choose a tag to compare

@github-actions github-actions released this 17 Jun 13:09

Changelog

New Features

  • ac254dc feat(ai): add a stop button to interrupt an in-progress chat response
  • 93f2f54 feat(ai): add copy button to chat messages
  • ae1ffb8 feat(ai): add diagnose_task_log tool for failure root-cause analysis
  • 1821ea2 feat(ai): add keyword + time-range filters to query_task_logs tool
  • b9e42f9 feat(ai): add list_templates tool so the chat can answer about task templates
  • d65d08d feat(ai): add offline search_docs tool over embedded docs
  • a1e0495 feat(ai): in-app AI ops chat panel with server-side tool-calling agent
  • b9e23cb feat(ai): make chat drawer width resizable by dragging its left edge
  • 96281d1 feat(ai): render assistant markdown in chat panel (zero-dep XSS-safe renderer)
  • 3c7241d feat(ai): stream chat over SSE with live tool-call display

Bug Fixes

  • 77ab3eb fix(ai): cap search_docs to one call to stop repeated slow re-searching
  • c16aa6d fix(ai): forbid fabricating HTTP API endpoints/fields/curl in answers
  • f73738d fix(ai): forbid inventing config-file sections; point notifications to the Web UI
  • dff9552 fix(ai): give chat prompt authoritative gocron facts (MCP, protocols, features)
  • 9b49615 fix(ai): instruct the model to reason in the user's language too
  • b3bb480 fix(ai): log tool calls, recover tool panics, surface max-iteration stalls
  • 86cd89f fix(ai): require user confirmation + audit before the chat runs a task
  • 826adb1 fix(ai): stop streaming being cut off mid-response by timeouts
  • e82e9a6 fix(ai): stop the model dangling a preamble without answering or calling tools
  • 03d0aaa fix(ai): stream reasoning_content so thinking models don't look stuck
  • 246d417 fix(ai): teach chat prompt gocron's 6-field cron and @-descriptors
  • ad2b814 fix(docs): patch vulnerable build deps via overrides; track docs/ in dependabot
  • b05dec2 fix(login): greet logged-in user by username instead of app name

Performance

  • 1bf93c4 perf(ai): analyze failures inline from query_task_logs, drop the extra LLM round-trip

Others

  • a19b8ae Merge docs into monorepo (VitePress site + AI docs)
  • 501e64c chore(docs): bring the VitePress docs site into the repo under docs/
  • 4abc1d7 chore: bump version to 1.6.5
  • 8e42450 docs(readme): mention the in-app AI ops chat in the AI Assist feature