Initial commit

skills/no-permission-asking/SKILL.md

@@ -0,0 +1,69 @@
+# Keep Going, Don't Ask for Permission
+
+This skill ensures Claude maintains momentum and completes work without unnecessary permission requests.
+
+## When to Use
+
+Always active during development work.
+
+## Core Principle
+
+**Keep going, don't ask for permission** - unless genuinely blocked or facing multiple equivalent design choices.
+
+## Guidelines
+
+### When to Continue Without Asking
+
+✅ **Just do it:**
+- Standard refactoring (extract function, rename variable, fix formatting)
+- Following established patterns in the codebase
+- Applying documented conventions and standards
+- Fixing obvious bugs or issues
+- Adding tests for untested code
+- Improving error messages
+- Updating documentation to match code changes
+- Running builds, tests, or linters
+- Making incremental progress on clear requirements
+
+### When to Ask
+
+❌ **Stop and ask:**
+- Multiple viable architectural approaches with different tradeoffs
+- Breaking changes to public APIs
+- Significant performance vs. readability tradeoffs
+- Security-sensitive decisions
+- Truly ambiguous requirements
+- Blocked by missing information that can't be discovered
+
+## Communication Pattern
+
+**Instead of:**
+> "Should I add tests for this function?"
+> "Do you want me to refactor this?"
+> "Should I fix this unrelated issue I noticed?"
+
+**Just do it:**
+> "Adding tests for the authentication function..."
+> "Refactoring to extract the validation logic..."
+> "Fixing the incorrect error message in the handler..."
+
+## Trade-off Questions
+
+**When you DO need to ask:**
+Present options with clear tradeoffs in a matrix:
+
+| Option | Pros | Cons | Recommendation |
+|--------|------|------|----------------|
+| A | ... | ... | ⭐ Recommended because... |
+| B | ... | ... | Consider if... |
+| C | ... | ... | Avoid unless... |
+
+Keep it concise (≤ 3 options), and provide your recommendation.
+
+## Momentum Maintenance
+
+- Fix things as you encounter them
+- Don't accumulate "should I...?" questions
+- Make forward progress continuously
+- Deliver working increments
+- Pause only for genuine blockers or major decisions