Prompt file imported from klintravis/CopilotCustomizer (
.github/prompts/Bootstrap.prompt.md). Fill in{{repositoryPath}}before use. Copyright stays with the author.
Bootstrap Workflow (v1.0)
Orchestrator: CopilotCustomizer.agent.md (routes to Bootstrap.agent.md)
Paired Instructions: Framework.instructions.md, RepoReview.instructions.md
Purpose
Fully autonomous workflow for generating Copilot customization assets for a target repository in the same workspace. Skills-first approach - prioritizes cross-platform capabilities (Skills) alongside VS Code-specific assets. Single entry point that orchestrates complete lifecycle: analysis → planning → generation → validation → harmonization → documentation.
Constraint: Do not target the CopilotCustomizer folder itself (use a different repository in the same workspace).
Skills-First Philosophy: Assets generated prioritize cross-platform Skills (agentskills.io standard) that work in VS Code, GitHub Copilot CLI, Claude, Cursor, and other compatible agents. Agents, Instructions, and Prompts are VS Code-specific and build upon the foundational Skills.
Usage
Minimal Input (One Command)
Bootstrap Copilot assets for: /path/to/repository
Variables
Repository Path [REQUIRED]: {{repositoryPath}}
That's it! Everything else is inferred automatically.
Workflow Overview
User Input (REPOSITORY_PATH)
↓
Bootstrap (validation)
↓
repo-analysis Skill (tech stack detection)
↓
Standards (match enterprise standards against tech stack)
↓
planning Skill (recommendations + specs + standards context)
↓
Orchestration Assessment (auto: lightweight-conductor if 3+ agents, orchestra/atlas for complex repos)
↓ [USER GATE: "confirm"]
Generator (create all assets - Skills PRIORITY, standards-informed)
↓
Verifier (schema validation + harmonization + metadata binding)
↓
documentation skill (report)
↓
COMPLETE
User Interactions: 2 (start + confirm plan)
Autonomous Phases: 7 (leveraging Skills for analysis, planning, documentation)
Quality Gates: 1 (after planning)
Skills Integration: repo-analysis, planning, documentation, deployment-automation (when applicable)
What Gets Inferred
From Repository Structure (using repo-analysis skill):
- Project type (API, CLI, web app, library)
- Tech stack (languages, frameworks, databases, testing)
- Existing patterns (architecture, conventions, deployment strategies)
- Current customization status
- Matched enterprise standards from
.github/standards/(if present)
Recommendations Generated (using planning skill):
- Skills (
.github/skills/) - Cross-platform capabilities [PRIORITY]- Generic Skills: repo-analysis, planning, asset-design, documentation, deployment-automation
- Project-specific Skills: e.g., api-development, react-patterns, rust-patterns
- Agent files (.agent.md) - VS Code specialists that leverage skills
- Instruction files (.instructions.md) - Coding standards for specific file patterns
- Prompt files (.prompt.md) - Task templates and slash commands
- AGENTS.md structure - Project guidance
- Cross-reference network - Metadata binding
Skills Are Portable: All generated Skills work across VS Code, GitHub Copilot CLI, Claude, Cursor, Goose, and compatible agents. See agentskills.io for full specification.
Example Execution
For a complete example walkthrough showing validation, analysis, planning, and autonomous execution phases, see EXAMPLES.md or QUICKSTART.md.
Validation Rules
Pre-Flight Checks:
Required:
- Repository path exists: PASS/FAIL
- Not CopilotCustomizer: PASS/FAIL
Recommended:
- Git repository: PASS/WARN
- VS Code workspace: PASS/WARN
Repository Exclusion:
if (repo.name.includes("CopilotCustomizer")) {
ABORT: "Target should be a separate repository in the same workspace.
Avoid running on CopilotCustomizer itself."
}
Agent Handoffs Schema (MANDATORY when present):
handoffs[] required fields per entry:
- label: string
- agent: string (must resolve to .github/agents/{agent}.agent.md)
- prompt: string (describes context transfer)
- send: boolean (true=auto, false=manual)
Validation Actions:
- Fail the run if any handoff entry is missing required fields
- Fail the run if
agentreference does not resolve to an existing agent file - Warn if
promptis too short or generic (< 20 chars)
Refinement Commands
After plan presentation, before confirmation:
refine: scope- Adjust asset countrefine: complexity- Simplify/expand specsadd: {asset}- Include additional assetskip: {asset}- Remove from planrefine: validation- Adjust strictness or report-only behavior for validation passescancel- Abort workflow
Error Scenarios
Invalid Repository Path:
Error: Repository not found at /invalid/path
Action: Please verify path and try again
CopilotCustomizer Detection:
Error: Target is the CopilotCustomizer folder
Action: Choose a different repository in the same workspace
Cannot Detect Tech Stack:
Warning: Unable to auto-detect project type
Action: Please specify: "Node.js API project with Jest testing"
Partial Generation Failure:
Warning: 6 of 8 assets created successfully
Failed: SecurityReviewer.agent.md, SecurityPatterns.instructions.md
Action: Continuing with available assets...
Validation Failure (Blocking):
Error: Handoffs schema validation failed
Details: Missing 'send' in 2 entries; unresolved agent reference: QAOrchestrator
Action: Aborting before documentation; fix generation specs and retry
Quality Guarantees
Acceptance Criteria:
- Repository path validated (not CopilotCustomizer)
- Tech stack detected (or manual input accepted)
- Asset recommendations generated
- <5 user interactions total (ideally 2)
- 90%+ handoff success rate
- Schema compliance: 100% (YAML + handoffs schema)
- Handoffs fields: 100% valid (label, agent, prompt, send)
- Verification + harmonization: PASS
- Cross-references: All resolved
- Documentation: Complete report
Performance Targets:
- Analysis: <60s
- Planning: <90s
- Generation: <120s per asset
- Validation: <30s
- Harmonization: <45s
- Documentation: <30s
- Total: <5 minutes for typical workflow
Framework Integration
Skills Used (cross-platform analysis and planning):
- repo-analysis - Tech stack detection, dependency analysis
- planning - Recommendation prioritization, strategy design
- documentation - Report generation and documentation
Framework Skills (available for all CopilotCustomizer workflows):
- asset-design - Asset decision framework
- deployment-automation - CI/CD and deployment strategies
- Example: GitHub Actions Pipeline
Reused Instructions (70%+ shared):
- Framework.instructions.md - Universal workflows
- RepoReview.instructions.md - Analysis patterns
- AgentAuthoring.instructions.md - Agent generation
- InstructionAuthoring.instructions.md - Instruction creation
- PromptAuthoring.instructions.md - Prompt templates
- SkillAuthoring.instructions.md - Skill creation (cross-platform)
- Maintenance.instructions.md - Asset optimization and harmonization
- Standards.instructions.md - Enterprise standards matching
Agent Chain (VS Code workflow orchestration):
- Bootstrap.agent.md - Entry point
- Planner.agent.md - Planning + gate (leverages planning skill)
- Generator.agent.md - Generation (prioritizes Skills)
- Verifier.agent.md - Validation + harmonization
- Editor.agent.md - File operations
Success Metrics
Efficiency:
- 70%+ instruction reuse from framework
- <5 user interactions per workflow
- <5 minutes total execution time
- 90%+ handoff success rate
Quality:
- 100% schema compliance
- 100% cross-reference resolution
- 0 orphaned assets
- Complete documentation
Autonomy:
- 1 input required (repository path)
- 1 approval required (confirm plan)
- 7 autonomous phases
- Zero manual intervention after confirmation
Framework: CopilotCustomizer v1.0
Workflow Pattern: Linear with single quality gate
Autonomy Level: Fully autonomous after confirmation
Powered by comprehensive CopilotCustomizer framework