Release v1.0.6 · ag2ai/ag2

What's Changed

fix(mcp): decode resource template URI variables by @asts-top in #3250
feat(mcp): let a served agent own its HTTP transport settings by @vvlrff in #3252
feat(tealtiger): add rate_limit governance policy by @nagasatish007 in #3246
feat(skills): render a MemorySkill body per read by @Lancetnik in #3254
fix(a2a): require only the chosen transport's SDK extra by @vvlrff in #3257
fix(subagents): lock max_concurrency semaphore provisioning per loop by @AmirF194 in #3243
fix(mcp): follow pagination when discovering tools by @asts-top in #3263
fix(anthropic): report thinking tokens, and keep a measured zero a zero by @vvlrff in #3259
feat(events): retain the provider's response identifier by @vvlrff in #3260
feat(openai): expose prompt cache controls for chat and responses by @vvlrff in #3261
feat(web-fetch): gate Anthropic-only options by version by @vvlrff in #3264
feat(web-fetch): control which sources supply fetchable URLs by @vvlrff in #3265
feat (OpenAI): prompt cache diagnostics by @vvlrff in #3266
refactor(mappers): build builtin tool payloads from the SDKs' param types by @vvlrff in #3270

Full Changelog: v1.0.5...v1.0.6