using-superpowerslisted

# Using Superpowers <SUBAGENT-STOP> If you were dispatched as a subagent to execute a specific task, skip this skill entirely. </SUBAGENT-STOP> ## Trigger Conditions This skill MUST be invoked when any of the following occur: - A new session starts with a technical request - The user gives a new task or changes topic mid-session - Any technical work is about to begin without a skill selected - The user asks "what should I use" or "which workflow" **Exception:** Micro tasks (typo fix, single variable rename, 1-line config change) can skip the entry sequence entirely. Just do them. ## When the User Names a Specific Skill If the user's prompt references a skill by name (e.g., "use brainstorming," "use context management," "run verification"), that is a **Skill tool invocation request**: 1. Still complete Entry Sequence steps 1–6 (token-efficiency, staleness check, etc.) — these are always-on prerequisites, not routing. 2. **Invoke the named skill via the `Skill` tool.** Do not re-implement the skill's purpose with ad-hoc agents, manual file reads, or improvised workflows. The skill contains tested, structured logic — use it. 3. Skip complexity classification and routing (step 7) — the user already chose the route. This is the most common cause of entry sequence bypass: the AI interprets "use X skill" as a goal to achieve creatively rather than as a tool invocation. It is always a tool invocation. ## Instruction Priority (highest to lowest) 1. Explicit user instruction