@fortressllm/sybil 0.0.3

package/docs/tutorials/skills.md

@@ -0,0 +1,561 @@
+# Dynamic Skills Tutorial
+
+Creating persistent learning capabilities for your Sybil agent.
+
+## Overview
+
+Skills are reusable capability modules that:
+- Learn from interactions
+- Provide context and instructions
+- Can be activated/deactivated per session
+- Persist across conversations
+- Follow Agent Skills specification `.md` format
+
+## Skill Structure
+
+Each skill is a `SKILL.md` file with this structure:
+
+```yaml
+---
+name: skill-name
+description: One-line description
+version: 1.0.0
+tags:
+  - category
+  - language
+  - topic
+---
+
+# Skill Title
+
+Instructions for when to use this skill
+
+## When to Use This Skill
+
+Clear scenarios of when to apply this skill
+
+## Guidelines
+
+Step-by-step instructions with examples
+
+## Examples
+
+Example 1: Usage scenario
+Example 2: Another scenario
+
+## Related Skills
+
+- related-skill-1
+- related-skill-2
+---
+*Generated by Sybil at 2026-02-12*
+```
+
+## Creating Skills
+
+### Method 1: `/create-skill` Command
+
+Create skills interactively:
+
+```
+User: Create a skill for managing GitHub pull requests
+```
+
+**Agent generates:**
+```yaml
+---
+name: pr-management
+description: GitHub PR management
+tags:
+  - github
+  - development
+version: 1.0.0
+---
+
+# GitHub Pull Request Management
+
+Expert at managing GitHub pull requests efficiently.
+
+## When to Use This Skill
+
+- User asks about PR operations
+- Need to open, merge, or comment on PRs
+- Workflow automation for PR reviews
+
+## Guidelines
+
+1. Always verify branch protections before suggesting operations
+2. Follow project-specific PR templates
+3. Consider context from pull request discussions
+4. Respect automated CI/CD pipelines
+5. Add value through meaningful comments and reviews
+
+## Examples
+
+Example 1: Merging PR
+User: Merge the hotfix changes
+```bash
+# Agent will check:
+# - Branch rules
+# - CI status
+# - Review status
+# # Merge if safe
+```
+
+Example 2: Commenting
+User: Comment on the recent security fix
+```bash
+# Agent will:
+# - Read PR description
+# - Generate contextual comment
+# - Post helpful technical feedback
+```
+
+## Related Skills
+
+- repository-management
+- code-review
+- ci-cd-automation
+---
+*Generated by Sybil at 2025-10-15*
+```
+
+### Method 2: `/network` Creation
+
+```
+User: Create a skill for handling Docker containers
+```
+
+**Network coordinator:**
+```
+[Researcher Agent]
+→ Researches Docker best practices
+→ Common commands, patterns
+
+[Executor Agent]
+→ Creates skill
+→ Writes code examples
+→ Includes error handling
+
+[Planner Agent]
+→ Structures skill into sections
+→ Adds learning objectives
+```
+
+---
+
+## Skill Activation
+
+### Manual Activation
+```
+User: /list-skills         # View all skills
+User: /activate-skill github-pr  # Activate skill
+```
+
+### Automatic Activation
+```
+Skills containing relevant keywords automatically activated based on:
+- User intent analysis
+- Previous interaction patterns
+- Contextual relevance
+```
+
+### Session-Based Activation
+```
+# Skills can be active for:
+- Single message: "activate-skill for this task only"
+- Current session: "keep active for this conversation"
+- Default: "enable for all conversations"
+```
+
+---
+
+## Skill Categories
+
+### Technical Skills
+- Development: pr-management, repository-management, debugging
+- Cloud: docker-containers, kubernetes, aws, azure
+- Security: vulnerability-scanning, penetration-testing
+
+### Business Skills
+- Communication: email-automation, schedule-automation
+- Sales: lead-scoring, CRM-management
+- Marketing: campaign-automation, analytics-reporting
+- Operations: incident-response, triage-workflow
+
+### Domains
+- Data Science: data-cleaning, etl-pipelines, visualization
+- Finance: risk-analysis, algorithmic-trading
+- Healthcare: patient-monitoring, medical-research
+- Education: course-generation, quiz-creation
+
+---
+
+## Learning Patterns
+
+### From Interactions
+
+**User Request:**
+```
+User: I'm tired of formatting dates by hand
+  User: Could you learn to do this?
+  Agent: I've detected pattern
+```
+
+**Agent Action:**
+```python
+# Creates skill: date-formatter
+# Includes: Date parsing, locale support, output formats
+```
+
+### From Feedback
+
+**User Feedback:**
+```
+User: The JSON format was hard to work with
+User: Could you learn a new output format?
+Agent: Updated skill-4 accordingly
+```
+
+### From Observation
+
+**Pattern Detected:**
+```
+Recurring: "Send me weekly reminders every Monday at 9am"
+→ Creates skill: weekly-reminder-scheduler
+```
+
+---
+
+## Version Management
+
+Skills support version updates:
+
+```yaml
+---
+version: 2.0.0
+---
+
+# Changelog
+
+v2.0.0:
+- Added timezone support
+- Added multiple reminders per week
+- Fixed DST handling
+
+v1.0.0:
+- Initial version with basic weekly reminders
+---
+```
+
+---
+
+## Managing Skills
+
+### View All Skills
+
+```
+User: /list-skills
+
+📚 Available Skills:
+  📊 github-pr-management (v1.0.0)
+  📅 weekly-reminder (v2.0.0)
+  🔧 docker-container-manager (v1.0.0)
+```
+
+### Activate Skills
+
+```
+User: Use the GitHub PR management skill
+→ Skill loaded into context
+→ All PR commands available
+```
+
+### Deactivate Skills
+
+```
+User: Stop using the reminder skill
+→ Skill deactivated for current session
+```
+
+### Update Skills
+
+```
+User: Update the docker skill to add volume management
+→ Creates updated version with version increment
+→ Old version archived for reference
+```
+
+### Delete Skills
+
+```
+User: Remove the old weekly-reminder skill
+→ Skill file deleted
+```
+
+---
+
+## Advanced Features
+
+### Skill Composing
+
+One skill can call another:
+
+```typescript
+ skills: ["base-skill", "advanced-skills"]
+```
+
+### Skill Context Injection
+
+When a skill is active, instructions are injected:
+```typescript
+instructions: `
+${SKILL_MD_PATH}
+${WORKING_MEMORY}
+---
+You have access to these skills: ${skills}
+`;
+```
+
+### Conditional Logic
+
+```typescript
+if (userContext.userType === "premium") {
+  skills.push("advanced-skills", "priority-support");
+}
+```
+
+### Skills with Tools
+
+Skills can specify tool dependencies:
+
+```typescript
+tools: ["browser-tools", "whatsapp-tools"];
+```
+
+---
+
+## Skill Performance Tracking
+
+### Metrics Collected
+- Usage frequency per skill
+- Success rate per task
+- User satisfaction ratings
+- Time saved per usage
+
+### Auto-Improvement
+
+Based on metrics:
+- ✅ Underutilized skills → archived
+- ✅ High-failure skills → flagged for review
+- ✅ Often-requested functionality → automatically creates new skills
+- ✅ Low-rated skills → offers improvements or alternatives
+
+---
+
+## Examples by Domain
+
+### Data Science Skills
+
+#### ETL Pipeline Skill
+```yaml
+---
+name: etl-pipeline-operator
+description: Manages ETL data pipelines efficiently
+tags:
+  - data-engineering
+  - etl
+  - pandas
+version: 1.0.0
+
+# ETL Pipeline Operations
+
+## When to Use This Skill
+
+When working with data extraction, transformation, and loading (ETL).
+
+## Guidelines
+
+1. Profile your workflow stages clearly
+2. Log each stage with metadata
+3. Monitor data quality at source
+4. Handle incremental vs full loads
+5. Document schema changes
+6. Implement data validation rules
+
+## Examples
+
+Example 1: Full Pipeline
+Pipeline: CSV → Transform → Load → DB
+
+Example 2: Incremental Load
+Pipeline: New rows → Validate → Append → DB
+
+Example 3: Schema Evolution
+Pipeline: Old schema → Transform to new schema → Backfill
+
+## Related Skills
+
+- data-validation
+- database-management
+- data-cleaning
+---
+```
+
+### DevOps Skills
+
+#### Kubernetes Deployment Skill
+```yaml
+---
+name: k8s-deployment-audit
+description: Audit and optimize Kubernetes deployments
+tags:
+  - kubernetes
+  - devops
+  - cloud
+version: 1.0.0
+
+# Kubernetes Deployment Auditing
+
+## When to Use This Skill
+
+When managing Kubernetes clusters for optimal performance and security.
+
+## Guidelines
+
+1. Audit resource requests and limits
+2. Check pod status and health
+3. Analyze logs for errors
+4. Review network policies
+5. Monitor storage and metrics
+6. Apply security best practices
+
+## Examples
+
+Example 1: Audit Deployment
+Audit: Check for misconfigurations
+
+Example 2: Optimize
+Action: Apply resource limits
+
+Example 3: Security Check
+Action: Scan vulnerabilities
+
+## Related Skills
+- docker-container-manager
+- cloud-security-audit
+- infrastructure-as-code
+---
+```
+
+### Skills with Tools
+
+### Automation Skills
+
+#### Automated Reporter
+Tools:
+- `browser-navigate`
+- `browser-get-content`
+- `browser-save-content`
+- `browser-execute-script`
+
+```yaml
+---
+name: automated-reporter
+description: Generate automated reports from dashboards
+tags:
+  - automation
+  - reporting
+version: 1.0.0
+
+# Automated Report Generation
+
+## When to Use This Skill
+
+When you need to generate periodic reports from web dashboards.
+
+## Guidelines
+
+1. Navigate to dashboard
+2. Wait for data to load
+3. Extract metrics
+4. Format as report
+5. Save or send report
+```
+
+### Skills for Multi-Agent Usage
+
+#### Research Coordinator
+**Skills:** [web-scraper, data-extraction, topic-synthesis]
+
+**Used by:** Researcher Agent
+
+---
+
+## Learning Curve
+
+### First 3 Skills
+
+1. Simple, domain-specific
+2. Easy to validate
+3. Clear success criteria
+
+### Beyond 3 Skills
+
+- Skills can depend on other skills
+- Complex workflows
+- Many configuration options
+- Requires more validation
+
+### Performance Tips
+
+- Skills with 1-2 parameters: ⚡ Fast activation
+- Skills with multiple branches: 🔀 More flexible
+- Skills with external APIs: 🌐 More info available
+- Skills with loops/iterators: 🔄 More testing needed
+
+---
+
+## Debugging Skills
+
+### Skill Not Working
+
+```typescript
+// Check skill file
+# Open: skills/my-skill/SKILL.md
+# Verify: Frontmatter is correct
+# Verify: Instructions are clear
+
+Common issues:
+- Missing fields in frontmatter
+- Invalid YAML syntax
+- Missing required sections
+- Conflicting examples
+```
+
+### Tool Not Available in Skill
+
+```typescript
+// Check tool imports
+// Include tools in skills directory
+
+"tools": {
+  "browser-tools": /* tools object */
+}
+```
+
+### Activation Issues
+
+```typescript
+// Check skill activation
+User: /list-skills
+→ Verify skill is active
+
+// Force activate specific skill
+User: /activate-skill my-skill
+→ Skill should load directly
+```
+
+---
+
+**Your agent now learns and grows with you! 🧠**