Project Context

Own the four regenerative project-context artifacts without duplicating their workflows across harness commands.

Compact Rules

• Exactly ONE mode per run: data · features · api · test-plan · refresh-all. Load only that mode's reference; never open a second one in the same pass.
• Mode → reference → output: data → references/data.md → .context/business/business-data-map.md · features → references/features.md → .context/business/business-feature-map.md · api → references/api.md → .context/business/business-api-map.md · test-plan → references/test-plan.md → .context/master-test-plan.md.
• User did not name a mode → ASK. NEVER infer refresh-all from a generic "refresh the context" request.
• refresh-all runs strictly data → features → api → test-plan, one at a time. Each reference's own validation and approval gate must close before the next is loaded. Never skip ahead.
• Artifact missing = CREATE mode: may write once the analysis completes. Artifact exists = UPDATE mode: generate a candidate, show the diff summary, WAIT for explicit approval. NEVER overwrite an existing artifact without that approval.
• Stop the run on a hard dependency failure or a rejected overwrite. A missing SOFT dependency is not a stop: record it as a Discovery Gap and continue, exactly as the selected reference defines.
• NEVER invent business facts. Read every source the selected reference requires; anything unverified belongs under the output's mandatory ## Discovery Gaps section, not asserted in the body.
• After a successful artifact write, add the pointer to AGENTS.md ONLY when that pointer is missing. Never add operational prose to CLAUDE.md.
• Forward $ARGUMENTS unchanged to the selected mode.

Read full SKILL.md when: the requested mode is ambiguous, a refresh-all chain fails mid-sequence, or you need the selected reference's own analysis steps and validation gate.

Mode routing