5W1H Decision Framework Tool. Use for: (1) Systematic decision-making before creating todos, (2) Preventing duplicate implementation, (3) Detecting avoidance behavior, (4) Ensuring agile refactor compliance with executor/dispatcher separation
Installation
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
5W1H-{TOKEN}
Who: {agent} (executor) | rosemary-project-manager (dispatcher)
What: {Single responsibility function}
When: {Event trigger with side effects}
Where: {Architecture layer / Component}
Why: {Requirement ID + Business value}
How: [Task Type: {TYPE}] {TDD strategy steps}
System-Level Enforcement
5W1H format is automatically enforced via:
Output Style (system prompt injection) - Always active
PreToolUse Hook - Validates todo creation
UserPromptSubmit Hook - Generates session token
Collaboration: when neurodivergent-output is also active
5W1H runs fully standalone. When the neurodivergent-output skill is also active, its cross-message state ledger becomes the persistent surface for decisions, and 5W1H adapts to the reader's cognitive-load rules. The composition triggers through that ledger (which fires every turn), so the compressed-5W1H rows ride on it rather than being a separate step to remember. Details:
Surface decisions in the ledger, compressed. When a 5W1H decision is made, put a compressed form — What, Why, next step — into the ledger's decisions section so it persists across messages. Do not dump the full six-field record, the session token, or the executor/dispatcher and Task-Type scaffolding into the ledger; keep the full record in its normal place and surface a pointer.
Respect the reader's load rules on the ledger. The ledger obeys neurodivergent-output's base layer (reduce overload, cap lists at five). Compress the 5W1H entry to fit; the ledger is a low-load anchor, not a full decision record.
Scope avoidance-language detection to decision content. Blocking "simplify / simpler / easier" applies to the work being decided, not to how the reply is shaped. Do not flag reader-facing output being compressed for an ADHD reader as avoidance.
Soften the gate under PDA. 5W1H's "all items must pass or BLOCKED" is a hard gate. When neurodivergent-output's PDA mode is on, present running 5W1H before a commit as an invitation with control left to the reader, not a block.
Verify it manifested — do not just declare it. When both skills are active, check each reply: are the ledger's decision rows really in compressed 5W1H form, or did only the ledger (the visible half) appear? The effortful half (5W1H structuring) silently drops if you rely on "both skills active" instead of checking the output. Declaration is not execution.
Neither skill depends on the other; this section only changes behavior when both are on.
