Anthropic engineer: Claude Tag writes >50% of my PRs every day
bcherny · x · 2026-09-26
Boris Cherny at Anthropic shares how he uses Claude Tag in Slack: it writes over 50% of his PRs daily, does 100% of his data analysis, and fixes most product feedback and bugs. Unlike a regular Slack bot, it's proactive, programmable, has memory, and can access your connectors — and with Opus 5.5 and Fable 5.1 it has strong judgment.
His example prompts include: react with ✅ in a channel once threads are resolved; automatically attempt end-to-end repro of every bug in a channel by running the full app, then put up a PR to fix it and tag the right team for review; brainstorm 100 hypotheses for weird data, use a workflow to validate them, spend 10M tokens digging deep, and draw a chart with the result; and build an interactive game explaining how a piece of code works, then a slide deck for the team.
Related event: Anthropic Engineer Says Claude Writes Over Half His PRs Daily(2 posts)→
More from coding & agent
- Perplexity CEO shows agent building browser-playable 3D zombie shooter in hours with zero 3D assets — AravSrinivas · 2026-11-11
- The complete harness engineering roadmap: how OpenAI shipped ~1M lines with zero hand-written code — blaizedsouza · 2026-11-11
- Micius-Flash: 27B Multimodal Model for Hardware and Semiconductor Agents — zhiman-ai · 2026-11-11
- Viral 'the harness is the machine' thread touts unverified Claude Fable/Mythos 5.2 9-equation architecture — blaizedsouza · 2026-11-11
- Compaction creates a new failure mode for coding agents: critical bug details can vanish mid-fix — blaizedsouza · 2026-11-11
- What 40+ coding-agent harnesses do between your prompt and the model, dissected — blaizedsouza · 2026-11-11