baa2d0220b
Add /ultraresearch-local for structured research combining local codebase analysis with external knowledge via parallel agent swarms. Produces research briefs with triangulation, confidence ratings, and source quality assessment. New command: /ultraresearch-local with modes --quick, --local, --external, --fg. New agents: research-orchestrator (opus), docs-researcher, community-researcher, security-researcher, contrarian-researcher, gemini-bridge (all sonnet). New template: research-brief-template.md. Integration: --research flag in /ultraplan-local accepts pre-built research briefs (up to 3), enriches the interview and exploration phases. Planning orchestrator cross-references brief findings during synthesis. Design principle: Context Engineering — right information to right agent at right time. Research briefs are structured artifacts in the pipeline: ultraresearch → brief → ultraplan --research → plan → ultraexecute. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
3.2 KiB
3.2 KiB
Knowledge Base Update Policy
Last updated: 2026-02 Applies to: ms-ai-architect plugin (5 skills, 364 reference files)
Update Frequency
| Priority | Category Pattern | Threshold | Rationale |
|---|---|---|---|
| Critical | cost, pricing, pris | 30 days | Azure prices change monthly |
| High | responsible-ai, norwegian-public-sector-governance, ai-security-engineering | 60 days | Regulations and compliance evolve quarterly |
| Medium | platforms, copilot-extensibility, azure-ai-services, multi-modal, performance-scalability, monitoring-observability, agent-orchestration, data-engineering, api-management, hybrid-edge, bcdr, rag-architecture, mlops-genaiops, prompt-engineering | 90 days | Feature updates follow Azure release cycles |
| Low | architecture, development, patterns | 180 days | Foundational patterns change slowly |
Category-to-Skill Mapping
| Category | Skill Directory | File Count |
|---|---|---|
| rag-architecture | ms-ai-engineering | ~20 |
| azure-ai-services | ms-ai-engineering | ~25 |
| copilot-extensibility | ms-ai-engineering | ~15 |
| prompt-engineering | ms-ai-engineering | ~15 |
| data-engineering | ms-ai-engineering | ~20 |
| api-management | ms-ai-engineering | ~10 |
| agent-orchestration | ms-ai-engineering | ~15 |
| multi-modal | ms-ai-engineering | ~10 |
| mlops-genaiops | ms-ai-engineering | ~15 |
| performance-scalability | ms-ai-engineering | ~10 |
| monitoring-observability | ms-ai-engineering | ~10 |
| responsible-ai | ms-ai-governance | ~25 |
| norwegian-public-sector-governance | ms-ai-governance | ~25 |
| cost-optimization | ms-ai-security | ~15 |
| ai-security-engineering | ms-ai-security | ~15 |
| hybrid-edge | ms-ai-infrastructure | ~15 |
| bcdr | ms-ai-infrastructure | ~15 |
| platforms | ms-ai-advisor | ~20 |
| architecture | ms-ai-advisor | ~20 |
Operational Procedure
Regular Check (Monthly)
-
Run staleness check:
bash scripts/kb-staleness-check.sh --json --output kb-status.json -
Review stale files by priority:
bash scripts/kb-staleness-check.sh --priority-only -
Update critical/high priority files:
/architect:generate-skills --update --priority critical /architect:generate-skills --update --priority high
Quarterly Review
- Run full staleness report
- Update all medium+ priority files
- Review and archive obsolete files
- Update this policy if thresholds need adjustment
Update vs Regenerate
- Update (preferred): Preserves existing structure, updates facts/dates/URLs. Uses Edit tool.
- Regenerate: Full rewrite. Use when file structure is outdated or content is >50% stale.
Quality Gates
- Updated files must pass:
bash tests/validate-plugin.sh - Updated files must have "Verified (MCP {month})" markers on MCP-sourced facts
- Updated files must maintain 7-15 KB size range
- No broken links or stale Microsoft Learn URLs
Automation
The SessionStart hook (session-start-context.mjs) automatically reports KB staleness levels at session start. The kb-staleness-check.sh script supports both human-readable and JSON output formats for integration with CI/CD or monitoring.