standard
Claude agent instruction file convention cited as a drop-in location for the completion contract paragraph.
No evidence-backed relationships are recorded.
Teams now maintain agent context files line by line because tokens are metered. The same teams left the human wiki to rot for a decade.
Reality
Anthropic says the model now verifies by default, so the double-check lines can go. The four-hundred-line diff behind the green check still has to be read by someone.
Reality
A developer avoided API-key coupling by shelling out to the Claude CLI. The default output format returns text only, so months of per-commit model calls left no usage record at all.
Reality
One team cut its agent context file to 34KB without deleting a single rule, then wired a size check and a structure test so it cannot quietly regrow.
Reality
A dependency rule for agent config files read only the scalar form of a YAML key. Authors who declared nothing went clean; authors who declared properly got the warning.
Reality
A thirty-run experiment finds agents followed long, buried rules anyway. The two failures came where the file contradicted the repository, not where it was long or buried.
Reality
An essay on dev.to argues that pasting runbooks and standards into every agent session is a platform capability nobody has claimed, and that the bill grows as AI adoption succeeds.
Reality
An agent reported a bulk insert complete when zero rows had landed. The fix is not a better guardrail on the text but a mandatory re-read of the system of record before completion can be claimed.
Reality