This release adds 2 notable features for engineering teams evaluating rollout.
✓ No known CVEs patched in this version
Topics
Summary
AI summaryDynamic LLM model updates and inline skill definitions improve crew wizard usability.
Full changelog
What's Changed
Features
- Pull latest LLM models dynamically in the crew wizard.
- Support inline skill definitions.
- Add generated Flow Definition authoring skill.
- Support templated Flow action inputs.
- Add text helper for flow CEL prompts.
- Add text helper to flow skill example.
- Implement message setup and feedback handling in AgentExecutor.
- Add repository agents to flow definitions.
- Define stream frame protocol for flows.
- Type tool and app in CrewDefinition.
- Repoint template commands to crewAIInc-fde org.
Bug Fixes
- Key model-catalog cache by exact API key, shorten TTL, and skip Ollama.
- Unify
crewai runflow input resolution and prompt from the state schema. - Resolve pip-audit failures for onnx 1.22.0 and nltk PYSEC-2026-597.
- Ensure we are writing version for flows.
- Include aiobotocore in the bedrock extra.
- Reject self-listening flow methods.
- Cut docs version nav from Edge so new pages aren't dropped.
Documentation
- Update language from Rules to Policies to match new dashboard changes.
- Document flow agent options.
- Add streaming docs to the navigation.
- Document Cost Limit rule type in Agent Control Plane.
- Drop CREWAI_LOG_FORMAT references from Datadog guide.
Contributors
@akaKuruma, @danielfsbarreto, @github-code-quality[bot], @joaomdmoura, @lorenzejay, @lucasgomide, @manisrinivasan2k1, @renatonitta, @vinibrsl
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About crewAI
Framework for orchestrating role-playing, autonomous AI agents. By fostering collaborative intelligence, CrewAI empowers agents to work together seamlessly, tackling complex tasks.
Related context
Related tools
Beta — feedback welcome: [email protected]