[TypeSpec Authoring] Integrate assessment with mitigation skills

## Goal

Allow assessment findings to be mitigated through the appropriate implementation skill, such as TypeSpec authoring or SDK breaking-change resolution.

## Scope

- Convert selected findings into a structured mitigation handoff containing the intent, impact, severity, source evidence, affected versions and APIs, and expected outcome.
- Route TypeSpec-source fixes to the TypeSpec authoring skill and generated SDK compatibility fixes to the applicable SDK breaking-change skill.
- Require user approval before editing code or invoking a mitigation workflow.
- Preserve links from the mitigation result back to the original findings.
- Reassess the changed code after mitigation and show which impacts were resolved, remain, or were introduced.
- Report clearly when no compatible mitigation skill is available; do not silently apply generic fixes.

## Acceptance criteria

- A user can select one or more findings and start the appropriate mitigation workflow.
- Skill handoffs use a documented, versioned input contract rather than report prose.
- Related findings with one root cause can be mitigated together without duplicate edits.
- Mitigation never changes code without explicit approval.
- Post-fix reassessment confirms the outcome and retains unresolved findings.
- Tests cover TypeSpec authoring, SDK breaking-change, unsupported-skill, declined-approval, partial-resolution, and failed-mitigation paths.

Start by locating the assessment, mitigation-skill, and reassessment entry points and their existing test suites; the issue names no files or commands. Trace how findings and skill handoffs are represented, then cover the listed TypeSpec, SDK, unsupported-skill, approval, partial-resolution, and failure paths, including preserved unresolved findings.

Written by the indexing model from the issue text.