v0.3.0

What's Changed

• feat(redteam): add built-in red teaming support by @kevmyung in #184
• chore: allow importing EvaluationReport from root by @poshinchen in #238
• chore: added trace-based evaluators into defaults by @poshinchen in #244
• chore(report): always return flattened report by @poshinchen in #241
• feat: add chaos resilience evaluators (failure communication, partial completion, recovery strategy) by @ybdarrenwang in #236
• fix(mappers): join all toolResult.content blocks to fix faithfulness false negatives by @poshinchen in #240
• chore(version): bumped strands-agents-version to the latest by @poshinchen in #246
• feat(redteam): add Crescendo multi-turn attack strategy by @yeomjiwonyeom in #245
• feat(cli): added strands-evals cli by @poshinchen in #243
• chore(evaluator): added evaluator name and evaluator_type for report by @poshinchen in #249
• chore(cli): added single case evaluation command by @poshinchen in #252
• fix(templates): correct doc link and clean up issue/PR templates by @yonib05 in #256
• docs: add AI contribution guidance to CONTRIBUTING and PR template by @yonib05 in #257
• docs(agents): add high-quality PR guidance to AGENTS.md by @yonib05 in #258
• docs(agents): add community and character guidance to AGENTS.md by @yonib05 in #261
• feat(issue-labeler): add LLM issue labeler for area and type by @yonib05 in #255
• feat(redteam): add Bad Likert Judge multi-turn attack strategy by @yeomjiwonyeom in #248
• chore(cli): added generate command for experiment generation by @poshinchen in #260

New Contributors

• @kevmyung made their first contribution in #184
• @yeomjiwonyeom made their first contribution in #245

Full Changelog: v0.2.1...v0.3.0