Research Artifacts

This directory contains pre-design research artifacts produced by 02-researching. Each artifact captures what was unknown before a design decision, the evidence used to resolve it, and the residual risks that downstream steps must account for.

When to write a research artifact

02-researching must materialize any resolved unknowns as a research artifact before handing off to the next step. If the research phase only confirmed facts that were already certain, no artifact is needed. If the phase resolved genuine unknowns, the artifact is mandatory.

Naming convention

Name each artifact <feature-slug>.md using the feature or decision slug from the active plan. Examples:

Required sections

Every research artifact must include the following sections:

Question

The precise unknown that the research phase was asked to resolve. One or two sentences. Link back to the originating plan step when possible.

Evidence

Source-grounded findings in strict order of authority: runtime/validation results, static source code, comments or documentation, then external references. For each finding, note the source and its strength. Record conflicting evidence explicitly with the tie-break rule used.

Decision

The decision that the evidence supports, stated as a single actionable conclusion. Include the rationale and any alternatives that were rejected.

Risks

Residual risks, open questions, or assumptions that remain after the decision. Name the owner phase or step that must validate or retire each risk.

Linking from a step packet

When a research artifact is produced, the 02-researching step packet must link to it. Use the research_artifact field in the step-packet YAML, or reference the artifact path in the handoff summary so the next step can locate the evidence.

Generated from source JSDoc • GitHub