Kjell Tore Guttormsen 0d0b83f98c feat: make examples cumulative with carry-forward chain and capstone

Add three new sections to all 14 examples:
- "Carry Forward": what output feeds into later examples (01-10)
- "The Cumulative Path": alternative prompt building on previous output (02-10)
- "Now Try It Yourself": personalized template with transferable pattern (all)
- "Building On" callout connecting back to previous examples (02-10)

Add Example 14: Build Your Personal Agent - capstone that guides reader
through writing their own CLAUDE.md, creating a personal skill, connecting
a messaging channel, setting up automation, and testing end-to-end.

Update README with cumulative path diagram, two usage modes, and example 14.
Update GETTING-STARTED.md with cross-references to relevant examples.

17 files changed, 703+ lines added. The examples now form a coherent
learning path from "see what it can do" to "build your own agent."

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

2026-03-26 21:14:35 +01:00

4.3 KiB

Raw Blame History

Example 06: Multi-Agent Orchestration

Capability: Claude Code can spawn sub-agents with distinct roles, run them in parallel or in sequence, and combine their outputs into a final result.

OpenClaw equivalent: Sub-agents, agent-to-agent messaging, mesh workflows.

Building on Examples 01-05. You have raw research (01), organized structure (02), verified data (03), and persistent state (05). Now three specialized agents turn that raw material into a polished, reviewed document. This is where the pipeline produces something you can actually share.

Prerequisites

This example uses the agents defined in .claude/agents/:

researcher.md - web research and source gathering
writer.md - structured content drafting
reviewer.md - accuracy and quality review

These agents load automatically when Claude Code opens this project.

The Prompt

Use the researcher agent to find information about how Claude Code handles
agent isolation and worktree sandboxing (introduced in v2.1.49 and v2.1.50).

Then use the writer agent to draft a 300-word technical summary of the findings,
written for a developer audience. No jargon without explanation.

Finally, use the reviewer agent to check the draft for technical accuracy.
If the reviewer finds any issues, have the writer fix them before showing
me the final version.

What Happens

Claude Code will:

Invoke the researcher agent via the Agent tool with the research task
Receive the research output and pass it to the writer agent
Invoke the writer agent to produce the 300-word draft
Invoke the reviewer agent with both the draft and source material
If the reviewer flags issues, loop back to the writer for a revision
Present the final reviewed draft

Why This Matters

Agent Teams (v2.1.32) gives Claude Code a mesh model that matches OpenClaw's sub-agent architecture. Worktree isolation (v2.1.50) means each agent gets its own working directory, preventing file conflicts in parallel runs.

The researcher-writer-reviewer pattern is the same loop that produces articles for fromaitochitta.com. The agents here are minimal versions of that pipeline.

Carry Forward

You now have a multi-agent review loop. This is the quality engine:

Example 07 delivers the reviewed output to your phone
Example 10 runs this exact agent sequence as steps 2-4 of the full pipeline
Example 14 shows you how to customize these agents for your own work

The researcher-writer-reviewer pattern is the single most reusable workflow in this repo. Any task that involves gathering information, drafting content, and checking quality follows this shape.

The Cumulative Path

If you ran Examples 01-05, you have a research report with verified data and persistent state. This prompt runs the full agent review cycle on it.

Read all files in pipeline-output/research-report/.

Use the researcher agent to verify any claims that have not been
web-verified yet and fill any gaps in the data.

Then use the writer agent to produce a polished 400-word summary
suitable for sharing with a colleague. Clear language, no jargon,
structured with headings.

Finally, use the reviewer agent to check for accuracy, clarity,
and completeness. If the reviewer finds issues, have the writer
fix them before showing me the final version.

Save the final version to pipeline-output/research-report/final-summary.md

After running this, your research has been through a professional review cycle. The final-summary.md is something you could email to your team.

Now Try It Yourself

Replace the demo topic with something you need researched and reviewed:

Use the researcher agent to find information about [your topic].
Then use the writer agent to draft a [length]-word [format] for
[your audience]. Use the reviewer agent to check [what matters
most: accuracy, tone, completeness]. Loop until the reviewer
approves.

The pattern you just learned: specialized agents + sequential handoff + revision loop. Break any complex task into roles (research, draft, review) and let agents handle each part.

Ideas worth trying:

Research a vendor and produce a one-page recommendation memo
Gather competitive intelligence and write an executive briefing
Compile technical documentation and review it for accuracy

4.3 KiB Raw Blame History