feat(projman): add plan-then-batch skill optimization

Separate cognitive work from mechanical API execution to reduce
skill-related token consumption by ~76-83% during sprint workflows.

Changes:
- Add batch-execution.md skill with 4-phase protocol
- Promote mcp-tools-reference and batch-execution to frontmatter
  for planner and orchestrator agents (auto-injected, zero re-read)
- Replace "Skills to Load" with phase-based "Skill Loading Protocol"
- Restructure planning-workflow.md Steps 8-10 for batch execution
- Update agent matrix in CLAUDE.md and docs/CONFIGURATION.md
- Add Phase-Based Skill Loading documentation section
- Clean up .gitignore (transient files, dev symlinks)

Token impact:
- 6-issue sprint planning: ~76% reduction
- 10-issue sprint planning: ~80% reduction
- 8-issue status updates: ~83% reduction

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

docs/CONFIGURATION.md

@@ -516,8 +516,8 @@ Agents specify their configuration in frontmatter using Claude Code's supported
 
 | Plugin | Agent | `model` | `permissionMode` | `disallowedTools` | `skills` |
 |--------|-------|---------|-------------------|--------------------|----------|
-| projman | planner | opus | default | — | body text (14) |
-| projman | orchestrator | sonnet | acceptEdits | — | body text (12) |
+| projman | planner | opus | default | — | frontmatter (2) + body text (12) |
+| projman | orchestrator | sonnet | acceptEdits | — | frontmatter (2) + body text (10) |
 | projman | executor | sonnet | bypassPermissions | — | frontmatter (7) |
 | projman | code-reviewer | opus | default | Write, Edit, MultiEdit | frontmatter (4) |
 | pr-review | coordinator | sonnet | plan | Write, Edit, MultiEdit | — |
@@ -591,6 +591,40 @@ The `skills` field auto-injects skill file contents into the agent's context win
 
 Skill names in frontmatter are resolved relative to the plugin's `skills/` directory. Use the filename without the `.md` extension.
 
+### Phase-Based Skill Loading (Body Text)
+
+For agents with 8+ skills, use **phase-based loading** in the agent body text. This structures skill reads into logical phases, with explicit instructions to read each skill exactly once.
+
+**Pattern:**
+
+```markdown
+## Skill Loading Protocol
+
+**Frontmatter skills (auto-injected, always available — DO NOT re-read these):**
+- `skill-a` — description
+- `skill-b` — description
+
+**Phase 1 skills — read ONCE at session start:**
+- skills/validation-skill.md
+- skills/safety-skill.md
+
+**Phase 2 skills — read ONCE when entering main work:**
+- skills/workflow-skill.md
+- skills/domain-skill.md
+
+**CRITICAL: Read each skill file exactly ONCE. Do NOT re-read skill files between MCP API calls.**
+```
+
+**Benefits:**
+- Frontmatter skills (always needed) are auto-injected — zero file read cost
+- Phase skills are read once at the appropriate time — not re-read per API call
+- `batch-execution` skill provides protocol for API-heavy phases
+- ~76-83% reduction in skill-related token consumption for typical sprints
+
+**Currently applied to:**
+- Planner agent: 2 frontmatter + 12 body text (3 phases)
+- Orchestrator agent: 2 frontmatter + 10 body text (2 phases)
+
 ---
 
 ## Automatic Validation Features