v0.0.6

What's Changed

• feat: include bounded failed-tool evidence in Jev decisions by @brianhong-dev in #10
• feat: record shadow tool recommendation outcomes by @brianhong-dev in #11
• feat: rerank large skill rosters with bounded detail by @brianhong-dev in #12
• feat: suggest a new approach after sustained low progress by @brianhong-dev in #13
• feat: mask configured values in Jev requests by @brianhong-dev in #14
• feat: qualify completion advice with observed evidence by @brianhong-dev in #15
• chore: prepare v0.0.6 release by @brianhong-dev in #16

Full Changelog: v0.0.5...v0.0.6