Addy on how to use Claude Code in the Claude Opus 5.5 era

Addy on how to use Claude Code in the Claude Opus 5.5 era

Starting with the Claude 5 series of models, the reliance on "stacking prompts and hand-holding task breakdowns" has declined; the more effective approach is a single prompt that clearly states the goal and how to verify it, letting the model run to completion on its own and then stop. The long system prompts written for weaker models in the old era, the forced splitting into sub-agents, and the manual monitoring of multiple sessions often add cost without necessarily making things smarter.

Prompts: from "teaching how to do it" to "what you want and how to verify it" For Opus 5.5, it is recommended to use a single prompt: clearly state what should be built, and how Claude should prove it has finished; then start tuning from /effort medium. When the goal is still unclear, first have Claude interview you until it can write a spec; then open a new session and use that spec to implement, rather than struggling along in the same noisy conversation.

Claude will stop on its own when it believes it is done. If acceptance is executable (pass/fail, a number to beat) and the report format is one it can read, you can let it continue in the background without having to watch it constantly. For long tasks, you can use /goal to set conditions, with another model checking after each round.

System prompts and "habits written for old models" For Claude Opus 5 and Claude Fable 5, more than 80% of the Claude Code system prompt was removed, with no obvious drop in coding benchmarks. After a new model comes out, you can run the same experiment on your own instructions: delete first, then add back only the parts the model actually misses. /doctor's prompt-audit will flag entries that still look like they were written for old models. It is recommended to audit your way of working every few months, to avoid still building processes for the old model era.

Permissions and auto mode Recently Claude Code defaults to auto mode in the interactive terminal: a classifier looks at the actions Claude wants to take, rather than popping up permission prompts everywhere. The author states clearly: this is not a security guarantee, and blindly approving everything is not either. When Claude reports results, you should have it show the commands and output, or screenshots.

Sub-agents: less manual splitting, more letting it dispatch on its own Claude will delegate to sub-agents on its own, and most of the time you don't need to name them. Only proactively request them for large, independently separable work such as long audits and migrations; Opus 5.5 can split this kind of work across parallel sub-agents. Forcing a split on trivial work raises cost and latency. Sub-agents can run in parallel and offer a second opinion, but by default they do not make the result smarter. When building your own agents, he observed: a single agent + skills often beats layers of sub-agents in quality, with lower cost and latency—consistent with his earlier view of "before adding sub-agents, first ask what problem they solve."

Cloud sessions, parallelism, and Projects To open a cloud session when you close your laptop (claude. ai/code, the Claude app, or claude --cloud): an independent virtual machine with a fresh clone of the repo, so push your local commits before starting; you can review on your phone, answer session questions, then review the diff and open a PR. When a task depends on resources unique to your local machine, keep the session local.

Parallel sessions are only worth it once a single session can run to completion on its own. Without Projects, a person has to act as the manager of multiple sessions. Projects (public beta on Pro / Max) is one long chat: issues / tasks pasted in each open their own thread (mostly cloud sessions) in parallel; unified instructions take effect once; Overview shows what is done, which PRs are ready, and what is waiting on you.

Review still matters, but automate a first read On the Claude Code team, most changes get at least one round of review and fixes from Claude before a human sees them. You can use Code Review on Team / Enterprise, or the lighter /code-review (reviewing the diff in a sub-agent that has not seen this conversation); you can also open a new session, give the spec, and say to ignore style issues.