Core Analysis¶

Goal: Run the MCP Server locally so VS Code can discover and install repository prompts/instructions/collections.

Basic Steps (Practical Guide)¶

1. Prepare environment: Install and start Docker; verify with docker run hello-world.
2. Run the MCP Server image: Example command from README:

docker run -i --rm ghcr.io/microsoft/mcp-dotnet-samples/awesome-copilot:latest

3. Configure VS Code: Add a stdio-type server in VS Code MCP settings with the above command.
4. Test discovery & install: Use / in Copilot Chat or the editor UI to search awesome-copilot and install a collection.

Common Issues & Fixes¶

• Docker permission/rootless: Ensure your user can run Docker or wrap the command in a startup script.
• Editor not recognizing stdio: Verify the command is non-interactive and the path is correct in MCP config.
• Fallback: If Docker is unavailable, copy repository files manually into local project structure.

Important Notice: MCP Server is a bridge; Copilot/Copilot Chat subscription and editor MCP support are still required.

Summary: Running the Docker image and configuring a stdio MCP server in VS Code is the most straightforward approach—provide a startup script and troubleshooting docs to ease adoption.

Core Analysis¶

Governance Goal: Balance openness with reliability to ensure prompts/instructions are high-quality, usable, and safe.

Recommended Governance Process¶

1. Contribution template & required frontmatter: Define fields (title, description, patterns, examples, risk-level) and require positive/negative example paths.
2. CI automated checks:
   - Markdown/YAML linting.
   - Pattern-matching tests in CI to validate patterns against sample files.
   - Content security scanning (detect leaked keys or dangerous commands).
3. Tiered review policy: Auto-merge low-risk items (docs/prompts), require manual approval for high-risk instructions that affect build/deploy/security.
4. Versioning & rollback: Use semantic versioning or changelogs for collections; enable quick rollbacks when problems arise.
5. Maintainer tooling & docs: Provide local validation scripts, pre-commit hooks, and contribution templates to improve consistency.

Caveats¶

• Enforce manual review and sample-run results for high-risk instructions.
• Track Copilot/Chat protocol changes and deprecate or update items as necessary.

Important Notice: Governance is rules plus a feedback loop (issues/reporting) for quick correction of misuse.

Summary: Standardize contribution flow and tie it to CI/review to transform community inputs into maintainable, auditable assets.

Core Analysis¶

Core Issue: File-pattern matching is essential for context-sensitive instructions but is the primary source of false positives (rules applied where they shouldn’t) and false negatives (rules not applied to intended files).

Common Error Types¶

• False positives: Overly broad globs (e.g. **/*.md) or failing to exclude third-party/generated directories.
• False negatives: Too narrow patterns, ignoring case sensitivity, alternate extensions, or path variants.
• Edge cases: Template/example files, generated code, or monorepo path differences.

Concrete Ways to Reduce Errors¶

1. Clear frontmatter syntax & examples: Include positive and negative path examples in each instruction file.
2. Whitelist + blacklist: Prefer explicit whitelists for intended targets and blacklist known noise dirs like node_modules or dist.
3. CI matching tests: Validate that new/edited patterns match expected sample files in PR checks.
4. Phased rollout: Enable rules on a limited set of files/modules first and collect feedback.

Important Notice: Do not rely solely on patterns—combine with code review, automated tests, and manual approval for high-risk rules.

Summary: Standardizing pattern authoring, example coverage, and CI validation reduces false positives/negatives to acceptable levels and makes instructions reliable in practice.