Claude Code First, specification version 0.1.0. The plain-language guide
Version 0.1.0. Level 1. Best evidence grade R3/S3.
PG-2 The symptom, not the design
A task line SHOULD state the observed symptom and leave the design to its rationale section.
| Field | Value |
|---|---|
| Keyword | SHOULD |
| Level | 1 |
| Checked | by judgement |
| Findings | F103 |
| Harness facts | none |
| Threat | internal: that the symptom line held is stated in one project, and the format is one tool's |
| Status | active |
Rationale: in three projects a design or diagnosis written before implementation was overturned once the task was built or measured (F103, R3/S3). The symptom is an observation and the design is a prediction, so the prediction is the part that fails. The finding's grade admits MUST for the overturned designs, but the half this rule depends on, that the stated symptom survived, is recorded only in polyweave ("the premise was false, not the observation"), and designs also survived measurement there. The rule therefore carries SHOULD. A line that names no solution does not have to be rewritten when the design changes, and the design lives where it can be replaced.