Recheck a newsletter analysis after the agent invents some posts
At a glance
Alesia Zakharova caught fabricated rows while Claude analyzed her published newsletter archive. This workflow is listed under data-verification, newsletter-analysis, browser-agents, failure-recovery.
She asked Claude for Chrome to compile metrics for fifty posts and compare topics. While reviewing its top-ten list, she noticed entries she had never written and asked it to rebuild the analysis from her actual archive.
How it went
She reports that Claude rebuilt the table and analysis after acknowledging that a truncated response had led it to fill gaps with invented post data.
Worth knowing
The author manually checked the corrected spreadsheet. The card does not treat later subscriber growth as proof that the initial analysis was reliable.
Save ideas in this browser. Upvotes count across the board.
An indie hacker packaged an agency's eight repeated workflows (briefs, client approval with revision counters, quotes with margin checks, content calendars, post-mortems) as free MIT Claude skills.
Anthropic’s Building Effective Agents appendix highlights two proven patterns—support agents that pull data and take actions like refunds, and coding agents that iterate against tests—with clear success metrics and human review.
Indie Hackers founder Sean Prentice used Claude Code to build an orchestrator + marketing director + specialist agents (content/ads/research) for his SaaS; outputs land in Slack and nothing posts or spends without his explicit say (~$10/mo self-use).