AI coding assistants fail to follow persistent instructions
🧭 workproduct problemSingle report ●●●●●
A developer reports that Claude Code consistently fails to adhere to project-specific guidelines defined in markdown files, leading to issues like context amnesia, repetitive errors, and the ignoring of established project rules.
What people do today
- Building custom enforcement hooks
- Writing repetitive markdown instructions
- Manual code review and correction
🔧 Fix this
Ideas a builder, maker or community group could start on:
- A standardized configuration schema for AI coding agents that enforces rules via system-level constraints rather than natural language prompts.
- A middleware tool that intercepts AI agent commands to validate them against a project-specific 'ruleset' before execution.
- A persistent memory management layer for coding agents that prevents context compression from discarding critical project instructions.
Real voices
“Claude consistently ignores CLAUDE.md - the very file Anthropic tells you to create. It s like hiring someone who forgets their job description every 2 hours.”
Related problems
🧭 workproduct19 reports
Indeed job feed ignores user preferences
🧭 workproduct14 reportsInability to filter AI-generated content on LinkedIn
🧭 workproduct11 reportsLinkedIn login verification loop
FixHumanityDaily
The most important new problems, and ideas to fix them, in your inbox.
Placeholder: no data is collected yet.