Simple CLAUDE.md tweak makes Claude Code output far easier to scan
julianharris · x · 2026-10-02
Developer julianharris shared a GitHub Gist showing how a formatting convention in CLAUDE.md makes Claude Code's terminal output dramatically more readable:
- Content addressed to the user (results, questions, decisions, approvals) gets wrapped in a consistent visual frame (🟦 first line + ▌ bars), so it stands out from the flood of progress logs and command narration.
- Progress and command output stay as plain one-liners outside the frame, preventing them from drowning out the actual replies.
- Line width is also limited to enable quick scanning.
He tagged the Claude developer account suggesting the readability improvement is worth adopting upstream. Readers can copy the Gist's CLAUDE.md snippet directly into their own projects to try it today.
More from coding & agent
- OpenAI and Perplexity rush out 'decision models' to compete with Jev, dev shares dead-code experiment — bendee983 · 2026-10-02
- Full-time agent developer: the hardest part of AI agents isn't AI, it's failure paths and handoff — jakes_takes_ · 2026-10-02
- remilouf teases 6-month agent project it calls a complete game changer — remilouf · 2026-10-02
- Polyphonic agents can autonomously live in the virtual world Somewhere — RileyRalmuto · 2026-10-02
- Researcher drafts an EMNLP 2026 A0 academic poster in HTML/CSS/JS — algo_diver · 2026-10-02
- Flask creator Armin Ronacher becomes chief MCP officer at Earendil, teases hot takes — mitsuhiko · 2026-10-02