add Serena gitignore and onboarding memories

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

.serena/.gitignore

@@ -0,0 +1,2 @@
+/cache
+/project.local.yml

.serena/memories/project_overview.md

@@ -0,0 +1,36 @@
+# Claude Skills Project Overview
+
+## Purpose
+A curated collection of agent skills following the [Agent Skills specification](https://agentskills.io/specification). Skills provide structured context (role, workflow, constraints, code examples) that AI coding agents load to perform specialized tasks. The project includes 66 skills across multiple domains, 9 workflow commands, and 365+ reference files.
+
+## Tech Stack
+- **Primary content**: Markdown (SKILL.md files, reference files, docs)
+- **Scripts**: Python 3 (validation, docs updates)
+- **Config**: YAML frontmatter in skill files, JSON (version.json, plugin.json)
+- **CI/CD**: GitHub Actions
+- **Linting**: ruff, ruff-format, pyright (Python); prettier (JS/JSON)
+- **Social preview**: Node.js + Puppeteer for screenshot generation
+
+## Project Structure
+```
+skills/                    # 66 skill directories
+  {skill-name}/
+    SKILL.md               # Tier 1: ~80-150 lines, frontmatter + body
+    references/             # Tier 2: 100-600 line deep-content files
+scripts/
+  validate-skills.py       # Validate YAML, descriptions, references, counts
+  update-docs.py           # Sync version/counts across all docs
+  validate-markdown.py     # Check markdown syntax issues
+docs/                      # Documentation site content
+commands/                  # 9 workflow command definitions
+.claude-plugin/            # Plugin manifest for Claude Code marketplace
+research/                  # Research docs (superpowers, etc.)
+version.json               # Single source of truth for version + counts
+CLAUDE.md                  # Project instructions for Claude
+MODELCLAUDE.md             # Model-facing behavioral instructions
+```
+
+## Key Design Principles
+- **Progressive Disclosure**: Metadata (~100 tokens) → SKILL.md (<5000 tokens) → References (on demand)
+- **Description Trap**: No process steps in descriptions. Format: `[Brief capability statement]. Use when [triggering conditions].`
+- **Framework Idiom Principle**: Reference files reflect idiomatic framework practices, not generic patterns

.serena/memories/style_and_conventions.md

@@ -0,0 +1,42 @@
+# Style and Conventions
+
+## Skill Frontmatter
+Required fields: `name`, `description`
+Optional: `license` (always MIT), `allowed-tools`, `metadata`
+
+### Description Format
+`[Brief capability statement]. Use when [triggering conditions].`
+- Capability verbs OK (generates, configures, implements)
+- Process steps NOT OK (first do X, then Y, finally Z)
+- Max 1024 characters
+- Must contain "Use when" somewhere in the description
+
+### Name Format
+- Lowercase letters, numbers, hyphens only
+- Must match directory name
+- Max 64 characters
+
+### Metadata Fields
+- `author`: GitHub profile URL
+- `version`: Quoted semver (e.g., "1.1.0")
+- `domain`: One of: language, backend, frontend, infrastructure, api-architecture, quality, devops, security, data-ml, platform, specialized, workflow
+- `triggers`: Comma-separated keywords
+- `role`: specialist | expert | architect | engineer
+- `scope`: implementation | review | design | system-design | testing | analysis | infrastructure | optimization | architecture
+- `output-format`: code | document | report | architecture | specification | schema | manifests | analysis | analysis-and-code | code+analysis
+- `related-skills`: Comma-separated existing skill directory names
+
+## SKILL.md Body
+- Target 80-150 lines (max 500)
+- Sections: Core Workflow, Reference Guide (routing table), Constraints (MUST DO / MUST NOT DO)
+- Inline code examples OK if under ~40 lines; larger examples go to reference files
+
+## Reference Files
+- 100-600 lines per file
+- Single topic focus
+- Complete working code examples
+- Cross-reference related skills
+
+## Python Scripts
+- Pre-commit hooks: ruff, ruff-format, pyright
+- Type hints required

.serena/memories/suggested_commands.md

@@ -0,0 +1,51 @@
+# Suggested Commands
+
+## Validation
+```bash
+# Full skill validation (YAML, descriptions, references, counts)
+python scripts/validate-skills.py
+
+# Single skill
+python scripts/validate-skills.py --skill react-expert
+
+# YAML checks only
+python scripts/validate-skills.py --check yaml
+
+# Reference checks only
+python scripts/validate-skills.py --check references
+
+# JSON output for CI
+python scripts/validate-skills.py --format json
+```
+
+## Markdown Validation
+```bash
+python scripts/validate-markdown.py
+python scripts/validate-markdown.py --path FILE    # Single file
+python scripts/validate-markdown.py --check        # CI mode
+```
+
+## Version & Docs Updates
+```bash
+# Update version in version.json first, then:
+python scripts/update-docs.py
+
+# Check files are in sync (CI)
+python scripts/update-docs.py --check
+
+# Preview changes
+python scripts/update-docs.py --dry-run
+```
+
+## Social Preview
+```bash
+npm install --no-save puppeteer && node ./assets/capture-screenshot.js
+```
+
+## Git / System
+```bash
+git status
+git log --oneline -10
+gh pr view NUMBER
+gh pr merge NUMBER --merge
+```

.serena/memories/task_completion_checklist.md

@@ -0,0 +1,22 @@
+# Task Completion Checklist
+
+When a task is completed, run the following before committing:
+
+1. **Validate skills**: `python scripts/validate-skills.py`
+   - Must pass with 0 errors (warnings OK)
+
+2. **Validate markdown** (if markdown files changed): `python scripts/validate-markdown.py`
+
+3. **Update docs** (if version/counts changed):
+   - Edit `version.json` with new version/counts
+   - Run `python scripts/update-docs.py`
+
+4. **Pre-commit hooks** run automatically on `git commit`:
+   - ruff (lint)
+   - ruff-format (format)
+   - prettier (JSON/JS)
+   - pyright (type check)
+
+5. **Changelog**: Update CHANGELOG.md under [Unreleased] or new version section
+
+6. **For releases**: Follow the full Release Checklist in CLAUDE.md