SEA-2.1 Completion Report — Skill Entity Dispatch Integration

**Files Changed:** - `[home-path]` (modified — added SEA skill entity detection + injection) - `[home-path]` (modified — added `--no-sea` flag + SEA status output) - **`detect_sea_skills(task, project_path)`** — Primary detection function that: - Tries SEA's embedding-based Tier 1 router first (via `sea_skill_injector.detect_sea_skills_for_task()`) - Falls back to keyword matching (`_sea_keyword_fallback()`) if Ollama/embeddings unavailable - Scans project's `CLAUDE.md` for explicit skill entity references - Returns enriched entity metadata (name, version, state, hot topics, confidence) - **`_load_skill_entity(skill_name)`** — Loads full entity context from: - `[home-path]` (description from YAML frontmatter) - `[home-path]` (hot topics, confidence, activations, mute status) - `[home-path]` (current version) - **`format_sea_skill_block(entities)`** — Formats detected entities into a prompt injection block with version, confidence, activation count, domains, and SKILL.md read paths. - SEA detection runs automatically on every dispatch (enabled by default) - SEA skills are deduplicated against the existing regex-based skill detection - Muted skills are filtered out - Added `enable_sea: bool` parameter (default `True`) - CLI supports `--no-sea` flag to disable