repositories / dotfiles

dotfiles

bugabingas dorkfiles

owned by bugabinga

pi: custom prompts (brainstorm, fix, refactor, review, test, tiger)

ed7bc24a1e81bb1bbc5b585d1f845f1f57dc18ec

Oliver Jan Krylow <oliver@bugabinga.net> ·

unsigned

Browse filesCompare with HEAD

Changed files (8)

pi/agent/prompts/brainstorm.md

Added
new file mode 100644
1 +---
2 +description: Brainstorm ideas
3 +argument-hint: "[focus-area]"
4 +---
6 +Brainstorm ideas for the project:
8 +**Focus**: $ARGUMENTS
10 +Use brainstorm skill phases:
12 +1. **Read Context** - Load AGENTS.md and explore code base
13 +2. **Diverge** - Generate many ideas wildly
14 +3. **Challenge** - Consult with the devil to find flaws
15 +4. **Defend** - Strengthen surviving ideas
16 +5. **Converge** - Pick best recommendations
18 +Brainstorm areas:
20 +- Features & capabilities
21 +- Architecture improvements
22 +- UX enhancements
23 +- Integration ideas
24 +- Performance optimizations
26 +Run all phases systematically. Ask clarifying questions. No fluff. Output ideas
27 +well formatted.

pi/agent/prompts/desloppy.md

Added
new file mode 100644
1 +---
2 +description: Find ugliest slop in code and fix it
3 +argument-hint: "[file|directory|scope]"
4 +---
6 +Find and fix sloppy code with no mercy:
8 +**Target**: $ARGUMENTS
10 +1. **Scan** - Find the ugliest code (complexity, duplication, poor patterns)
11 +2. **Identify** - What makes it slop? (nesting, long functions, magic numbers,
12 + no types)
13 +3. **Fix** - Rewrite cleanly
14 +4. **Verify** - Run test after fixes
16 +Focus on:
18 +- Deep nesting → flatten
19 +- Long functions → extract
20 +- Magic values → name them
21 +- Missing types/docs → add them
22 +- Poor error handling → fix it
23 +- overengineered error handling -> simplify
24 +- overengineering due to "legacy" or "backwards compatibility" -> verify the
25 + need
27 +No mercy. Make it beautiful. Report first, wait for approval.

pi/agent/prompts/fix.md

Added
new file mode 100644
1 +---
2 +description: Fix bug with systematic debugging
3 +argument-hint: "<bug-description-or-error>"
4 +---
5 +Debug and fix the issue:
7 +**Problem**: $ARGUMENTS
9 +1. **Reproduce** - Get exact steps or error message
10 +2. **Hypothesize** - List possible causes
11 +3. **Investigate** - Search code, check logs, examine state
12 +4. **Test Fix** - Apply minimal fix
13 +5. **Verify** - Ensure fix works, no regressions
14 +6. **Prevent** - Add test to catch future regressions

pi/agent/prompts/optimize.md

Added
new file mode 100644
1 +---
2 +description: Benchmark and optimize code
3 +argument-hint: "<function-or-file>"
4 +---
6 +Benchmark and optimize:
8 +**Target**: $ARGUMENTS
10 +Workflow:
12 +1. **Profile** - Use profiling to find hotspots and other performance problems
13 +2. **Baseline** - Create benchmark
14 +3. **Analyze** - Check code
15 +4. **Optimize** - Write efficient code
16 +5. **Verify** - Re-run benchmark, compare with benchstat
17 +6. **Document** - Comment optimization rationale

pi/agent/prompts/refactor.md

Added
new file mode 100644
1 +---
2 +description: Refactor code safely
3 +argument-hint: "<file-or-directory>"
4 +---
6 +Refactor code:
8 +**Target**: $ARGUMENTS
10 +1. **Analyze** - Understand current structure
11 +2. **Plan** - List refactoring steps
12 +3. **Tests** - Ensure coverage before changes (use /test)
13 +4. **Refactor** - Apply changes incrementally
14 +5. **Verify** - Run tests after each step
16 +Common patterns: extract function, remove duplication, simplify conditionals.

pi/agent/prompts/review.md

Added
new file mode 100644
1 +---
2 +description: Quick code review with checklist
3 +argument-hint: "[file-or-directory]"
4 +---
6 +Review code using systematic checklist:
8 +**Target**: $ARGUMENTS
10 +1. **Correctness** - Logic errors, edge cases, off-by-one
11 +2. **Security** - Injection, auth, secrets, input validation
12 +3. **Performance** - O(n²) traps, unnecessary allocs, N+1 queries
13 +4. **Readability** - Naming, complexity, comments
14 +5. **Testing** - Coverage gaps, missing edge cases
16 +Provide specific line references and fix suggestions.

pi/agent/prompts/test.md

Added
new file mode 100644
1 +---
2 +description: Generate comprehensive tests
3 +argument-hint: "<file-or-function>"
4 +---
6 +Generate tests:
8 +**Target**: $ARGUMENTS
10 +1. **Unit Tests** - Happy path, edge cases, error cases
11 +2. **Integration** - If interacts with external systems
12 +3. **Snapshot** - Use snapshot testing for complex output
14 +Aim for 100% coverage on critical paths.

pi/agent/prompts/tiger.md

Added
new file mode 100644
1 +---
2 +description: Enforce tiger style on code
3 +argument-hint: "[file-or-directory]"
4 +---
5 +Enforce tiger style with no mercy:
7 +**Target**: $ARGUMENTS
9 +Use tiger skill to scan and enforce:
10 +1. **Scan** - Detect non-tiger patterns
11 +2. **Enforce** - Apply tiger style rules
12 +3. **Rewrite** - Transform code to tiger standards
14 +Tiger style principles:
15 +- Readable, maintainable code
16 +- Proper naming conventions
17 +- Consistent formatting
18 +- Clear structure
19 +- No unnecessary complexity
21 +No mercy. Tiger style or death.