## Step 1: Read the report Read the issue or report fully. Identify what the user expected, what happened, and any reproduction steps. ## Step 2: Classify - **Won't do** — conflicts with project goals, out of scope, or intentional behavior - **Confirmed bug** — unexpected behavior with a clear reproduction path - **Clear win feature** — aligns with REQUIREMENTS.md, limited scope, adds obvious value - **Docs gap** — behavior is correct but undocumented or misleadingly documented - **Unclear** — cannot classify without more information To classify correctly: - Read `REQUIREMENTS.md` (Must Have / Should Avoid / Out of Scope) - Read `CLAUDE.md` — many apparent bugs are documented non-obvious behaviors - Search the source if needed to confirm whether the behavior is by design ## Step 3: Resolve **Won't do:** Answer the user directly. Explain why, referencing `REQUIREMENTS.md` if helpful. **Confirmed bug:** Write a task in `TODO.md`: - Next available `CE-NNN` ID - **Skill:** `cause-effect-dev` - **Context:** reproduction steps and expected behavior **Clear win feature:** Confirm alignment with `REQUIREMENTS.md`. Write a task in `TODO.md`: - **Skill:** `cause-effect-dev` - **Context:** what to add and why it aligns with project goals **Docs gap:** Write a task in `TODO.md`: - **Skill:** `tech-writer` - **Context:** which document is unclear and what the correct behavior is **Unclear:** Ask the user. Do not write a task or make assumptions. - Issue classified with clear reasoning grounded in REQUIREMENTS.md or CLAUDE.md - Won't do: user has a specific explanation - All other resolvable cases: a correctly formatted task in TODO.md - Unclear cases: escalated to the user, not guessed