v1.2.0 — Autonomous Agent System
Autonomous Agent System
Research-heavy commands now delegate to autonomous agents that run in isolated context windows, keeping web research context separate from the main conversation.
New Agents
| Agent | Command | Purpose |
|---|---|---|
arckit-research |
/arckit.research |
Market research, vendor evaluation, build vs buy, TCO |
arckit-datascout |
/arckit.datascout |
Data source discovery, API catalogue search, scoring |
arckit-aws-research |
/arckit.aws-research |
AWS service research via AWS Knowledge MCP |
arckit-azure-research |
/arckit.azure-research |
Azure service research via Microsoft Learn MCP |
How It Works
- Slash commands become thin wrappers that launch agents via the Task tool
- Agents run autonomously in their own context window (dozens of WebSearch/WebFetch/MCP calls)
- Results are written to file; only a summary returns to the main conversation
- Fallback to direct execution if agent delegation fails
- Claude Code only — Codex CLI and Gemini CLI retain full inline prompts
Files Added
.claude/agents/arckit-research.md.claude/agents/arckit-datascout.md.claude/agents/arckit-aws-research.md.claude/agents/arckit-azure-research.md.claude/settings.json
Documentation Fixes
- Fixed missing example links for 6 commands (stakeholders, risk, sobc, azure-research, aws-research, gcloud-search)
- Fixed broken platform-design v8 link
- Corrected Wardley Maps, ServiceNow, and Diagrams prose to reference repos that actually contain those artifacts
- Updated Supported AI Agents section to document agent architecture
Full Changelog
See CHANGELOG.md for details.