B-TIER resource
CLAUDE.md Workflow Template — plan mode, subagents and a lessons loop
100k ft — What is this?
CLAUDE.md Workflow Template — plan mode, subagents and a lessons loop
50k ft — Why care?
The CLAUDE.md template attributed to Boris Cherny's workflow: plan mode for any task of 3+ steps, liberal subagent use, a tasks/lessons.md file updated after every correction, verification before marking done, 'demand elegance', and autonomous bug fixing. The caption adds that he runs 10-15 Claude sessions in parallel across terminal and web; the same template also closes the fuller tips thread in this batch.
10k ft — Key takeaways
• Never mark a task done without proving it works, for example by diffing behavior against main.
• Use subagents to keep the main context window clean.
• Write the plan to tasks/todo.md with checkable items and log corrections to tasks/lessons.md.
5k ft — Tools & techniques
Anthropic, Claude, Claude Code
CLAUDE.md rules, parallel sessions, plan mode, rule-ratchet CLAUDE.md, self-improvement loop, subagents, workflow automation
1k ft — Deep-dive / try it
Boris Cherny, the engineer behind Anthropic’s Claude Code, has outlined a high-throughput workflow built around running 10–15 Claude sessions in parallel—split between terminal and web—to ship multiple workstreams at once.
He says the real leverage comes from maintaining a CLAUDE.md “rules” file that is continuously updated after mistakes, effectively turning corrections into reusable guardrails for the next run.
Cherny also described using Claude to pull data through the CLI rather than hand-writing queries, and said Claude Code has grown to a meaningful share of public GitHub activity
julienbaneux
18w
Anyone actually try these yet?