your-skill-namelisted

# Your Skill Name ## Overview What this is. The core principle in 1-2 sentences — the one thing an agent should remember if it forgets everything else. ## When to Use - Symptom or situation that should trigger this skill - Another trigger - Another trigger **When NOT to use:** - Adjacent case that looks similar but isn't - Another exclusion ## The Workflow 1. Step with its purpose 2. Step — include the decision point 3. Step <!-- If there is a genuine decision loop, a small flowchart beats prose: ```dot digraph d { "Question?" -> "Path A" [label="condition"]; "Question?" -> "Path B" [label="other condition"]; } ``` --> ## Hard Rules - Rule stated as a prohibition, with the reason it exists (the scar) - Rule ## Rationalizations — STOP | Excuse | Reality | |---|---| | "Quote the excuse agents actually make" | "Why it's wrong — concretely" | ## Red Flags — STOP - Self-checkable behavior that means the agent is about to violate the skill - "This case is different because..." ## Quick Reference <!-- Heavy operational reference (commands, APIs, long tables) belongs in a supporting file next to this SKILL.md, e.g. templates.md, linked like: See [templates.md](templates.md). Load when executing, not when deciding. Create that file before shipping — CI enforces that every relative link resolves. -->