🌟 Release Highlights
This is a focused patch release that resolves a long-standing step summary truncation issue affecting all AI engines, alongside a documentation improvement and an internal test reliability fix.
🐛 Bug Fixes & Improvements
- Step summary output expanded to 2000 characters — Agent step summaries were silently truncated at 500 characters. This release increases the limit to 2000 characters and correctly forwards
GITHUB_STEP_SUMMARYinto the sandbox for all engines (Copilot, Codex, Claude, Gemini), ensuring agents can write meaningful summaries without silent data loss. (#19821)
📚 Documentation
- Added a PDF download link to the slides documentation page, making it easier to share and reference presentation materials offline. (#19842)
🌍 Community Contributions
A huge thank you to the community members who reported issues that were resolved in this release:
@danielmeppielfor Step summary truncates agent output at 500 chars with no visible warning (#19810)
For complete details, see CHANGELOG.
Generated by Release
What's Changed
- Add PDF download link to slides doc page by @Copilot in #19842
- Fix label trigger integration test type assertions for
namesfield by @Copilot in #19843 - Fix step summary truncation: forward GITHUB_STEP_SUMMARY into sandbox for all engines, increase text limit to 2000 chars by @Copilot in #19821
Full Changelog: v0.53.5...v0.53.6