Subagent prompts should include explicit instructions to read CLAUDE.md

[0xDones]

Describe the bug
I have very clear conventions written to CLAUDE.md file, but every plan/implementation is not following it, I think it would be good to have explicit instructions on the subagent prompt to read it, or just provide it as additional context.

To Reproduce
Steps to reproduce the behavior:

1. Have well defined instructions on CLAUDE.md
2. Write a plan and execute it

Expected behavior
A clear and concise description of what you expected to happen.

1. Should provide context from CLAUDE.md file to subagents

Logs
If applicable, add logs to help explain your problem.

Additional context
Add any other context about the problem here.

[Koroqe]

Had this exact same frustration — CLAUDE.md rules getting ignored by subagents. What actually worked for me was moving away from relying on instruction-following entirely and instead building the rules into the workflow structure itself. Each agent in the pipeline only sees what it needs and has to produce specific artifacts before the next step runs. Basically shifted from "please follow these rules" to "you literally can't skip this step." Been using this approach: https://github.com/Koroqe/claude-code-sdlc — the enforcement comes from the pipeline gates, not from hoping subagents read the docs.