Align Cockpit release lens with SDLC state helpers

[smota]

Summary

Aligns Cockpit release lens derivation with shared SDLC state helpers so Cockpit consumes the SDLC product model instead of maintaining separate release candidate/impact/assignment logic.

Implemented issues

Implements #171
Refs #163

Related issues

Parent production epic: #163. Prior slices implemented #164-#170 and #172.

Workflow evidence

Workflow-status comment: exception: PR body records this alignment slice.

Handover comments: exception: release train continues in current session toward #163 closeout.

Role-pass summary:

• Updated lib/cockpit-goal-model.mjs to import releaseCandidateFromIssue, releaseImpactFromIssue, and releaseAssignmentState from lib/sdlc-state.mjs.
• Removed duplicate Cockpit-local release candidate and impact helper logic.
• Preserved Cockpit behavior while centralizing SDLC release rules.

Validation evidence:

• pnpm test passed: 22 files / 144 tests.

Agent review

• Implemented by: pi
• Launcher: pi
• Executor: pi-parent
• Transport: local-cli
• Delegation boundary: current-session
• Model / runtime: gpt-5.5 / pi
• Review: self-review
• Workflow profile: standard
• Merge owner: human/operator
• Mode: single-agent
• Self-review disclosure: Same session implemented and reviewed; focused code change covered by existing Cockpit + SDLC tests.

CI-equivalent validation

• Status: passed
  Commands:
• pnpm prettier --write lib/cockpit-goal-model.mjs
• pnpm test

Follow-up issues

• Epic: AgentFlow SDLC Definition Product v1 #163 production release closeout remains parent tracker.

Out of scope

• Docker packaging.
• Additional Cockpit UI changes.