v3.0.0

What's Changed

• feat: [V3 PM] Define the human and AI PM operating policy and harness contract by @StatPan in #870
• feat: [V3 PM] Add pm-ir/v1 and read-only intent compiler diagnostics by @StatPan in #872
• feat: [V3 PM] Add typed PM context, evidence, assumption, and decision ledger by @StatPan in #873
• feat: [V3 PM] Add compact typed task profiles and profile-aware readiness by @StatPan in #874
• feat: [V3 PM] Model discovery, hypotheses, experiments, outcomes, and learning by @StatPan in #875
• feat: [V3 PM] Add outcome-aware measurement plans and validation by @StatPan in #876
• feat: [V3 PM] Compile typed Goal work graphs into verifiable child issues by @StatPan in #878
• feat: [V3 PM] Add the active PM observe and fingerprint-guarded replan loop by @StatPan in #879
• feat: [V3 PM] Persist PM acceptance, outcome validation, and learning transitions by @StatPan in #880
• feat: [V3 PM] Derive PM reports and human/AI handoffs from canonical state by @StatPan in #881
• feat: [V3 PM] Ship CLI/MCP PM harness parity and conformance benchmarks by @StatPan in #883
• feat: Honor recorded work_branch when validating PR bindings by @StatPan in #885
• feat: Prevent hidden merge expansion after Draft finish preview by @StatPan in #886
• feat: Preserve merged PR binding across finish resolution and receipts by @StatPan in #887
• feat: Add opt-in visual portfolio report for milestones, timeline, and work queues by @StatPan in #888
• feat: Prepare v3.0.0 changelog and release audit by @StatPan in #891

Full Changelog: v2.8.0...v3.0.0