We split this CLAUDE.md into the rules it gives an agent

9 rules. About 0 are common, 3 are distinctive, and 6 are very specific to this project.

• ≈0

  Common · estimate

  Many files like this have a rule of this kind.

• ≈3

  Distinctive · estimate

  This author's own choices. Worth a look.

• 6

  Very specific

  Tied to their project's names, files or tools. Probably skip.

• Common
• Distinctive
• Very specific

1. Very specific:

   Before starting any session, read and follow .aicontext/prompts/start.md.

   line 3
2. Distinctive:

   When subagents are available, delegate these tasks instead of performing them inline:

   line 7
3. Very specific:

   NEVER perform a review inline when the reviewer agent exists — delegate to the agent

   line 19
4. Distinctive:

   NEVER read large amounts of code into main context for review — delegate to the agent

   line 20
5. Very specific:

   Run multiple agents in parallel when tasks are independent (e.g., reviewer + test-runner after implementation)

   line 21
6. Very specific:

   NEVER use EnterPlanMode — all planning is done in .aicontext/tasks/ task files per process.md

   line 22
7. Very specific:

   NEVER launch a subagent without telling the user first — state which agent and what it will do (e.g., "Asking subagent test-runner to …

   line 23
8. Distinctive:

   When invoking an agent, provide only the context it needs (command, directory, files, task requirements).

   line 24
9. Very specific:

   When a skill is invoked via /command, its content is provided inline — NEVER read files from .claude/skills/ directories

   line 25

≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.