opencode-swarm-plugin 0.12.27 → 0.12.28

package/global-skills/code-review/SKILL.md

@@ -1,166 +0,0 @@
----
-name: code-review
-description: Code review patterns for agents. Use when reviewing changes before committing, evaluating pull requests, or assessing code quality. Provides systematic review checklists and feedback patterns.
-tags:
-  - review
-  - quality
-  - best-practices
-tools:
-  - swarm_evaluation_prompt
----
-
-# Code Review Skill
-
-Systematic code review patterns for ensuring quality.
-
-## When to Use
-
-- Before committing changes
-- Reviewing pull requests
-- Evaluating code quality
-- Self-review before completing subtasks
-- Assessing unfamiliar code
-
-## Review Protocol
-
-### 1. Understand Context
-
-Before reviewing code:
-- What is this change trying to accomplish?
-- What files/modules are affected?
-- Are there related tests?
-- Is this a bugfix, feature, or refactor?
-
-### 2. Quick Scan
-
-First pass - high level issues:
-- Does the change make sense overall?
-- Is the approach reasonable?
-- Any obvious red flags?
-
-### 3. Detailed Review
-
-For each file changed:
-
-**Correctness**
-- Does the logic do what it's supposed to?
-- Are edge cases handled?
-- Are there potential bugs?
-
-**Security**
-- Input validation?
-- Authentication/authorization checks?
-- Sensitive data exposure?
-- Injection vulnerabilities?
-
-**Performance**
-- Unnecessary loops or database calls?
-- Memory leaks potential?
-- Appropriate data structures?
-
-**Maintainability**
-- Clear naming?
-- Reasonable complexity?
-- Follows existing patterns?
-
-**Testing**
-- Tests for new functionality?
-- Edge cases tested?
-- Tests actually test the right things?
-
-### 4. Provide Feedback
-
-Structure feedback clearly:
-
-```markdown
-## Summary
-[Overall assessment - approve, needs changes, questions]
-
-## Issues
-- [Critical] Issue description and suggestion
-- [Minor] Style issue
-
-## Questions
-- Why was X approached this way?
-
-## Suggestions (optional)
-- Consider using Y pattern here
-```
-
-## Review Checklists
-
-### For New Features
-
-- [ ] Feature works as described
-- [ ] Tests cover happy path and edge cases
-- [ ] Error handling is appropriate
-- [ ] No hardcoded values that should be configurable
-- [ ] Documentation updated if needed
-- [ ] No TODO comments for critical functionality
-
-### For Bug Fixes
-
-- [ ] Bug is actually fixed
-- [ ] Regression test added
-- [ ] Root cause addressed (not just symptoms)
-- [ ] Similar issues elsewhere checked
-- [ ] No new bugs introduced
-
-### For Refactoring
-
-- [ ] Behavior unchanged (all tests pass)
-- [ ] Code actually improved
-- [ ] No unnecessary changes
-- [ ] No feature changes mixed in
-
-## Self-Review Before Commit
-
-Use `swarm_evaluation_prompt` for structured self-evaluation:
-
-```
-swarm_evaluation_prompt(
-  bead_id: "bd-xxx",
-  subtask_title: "What I implemented",
-  files_touched: ["file1.ts", "file2.ts"]
-)
-```
-
-Criteria:
-- **type_safe**: Does it compile without errors?
-- **no_bugs**: Any obvious bugs or edge cases?
-- **patterns**: Follows existing code style?
-- **readable**: Would another developer understand it?
-
-## Common Issues to Catch
-
-### Logic Errors
-- Off-by-one errors
-- Incorrect boolean logic
-- Missing null checks
-- Race conditions
-
-### Style Issues
-- Inconsistent naming
-- Magic numbers
-- Deep nesting
-- Large functions
-
-### Security Issues
-- SQL injection potential
-- XSS vulnerabilities
-- Hardcoded credentials
-- Missing auth checks
-
-### Performance Issues
-- N+1 queries
-- Unnecessary re-renders
-- Memory leaks
-- Blocking operations
-
-## Swarm Integration
-
-When reviewing in a swarm:
-1. **Coordinate reviews** - Don't duplicate effort
-2. **Share patterns** - If you catch something, broadcast it
-3. **Track issues** - Create beads for non-blocking issues
-4. **Learn together** - Good review findings become skills

package/global-skills/debugging/SKILL.md

@@ -1,150 +0,0 @@
----
-name: debugging
-description: Systematic debugging patterns for agents. Use when encountering errors, unexpected behavior, or when tests fail. Provides structured approaches for root cause analysis and error resolution.
-tags:
-  - debugging
-  - errors
-  - troubleshooting
-tools:
-  - swarm_accumulate_error
-  - swarm_get_error_context
-  - beads_create
-  - semantic-memory_store
----
-
-# Debugging Skill
-
-Systematic approaches for diagnosing and resolving issues.
-
-## When to Use
-
-- Encountering error messages or exceptions
-- Tests failing unexpectedly
-- Behavior differs from expectations
-- Performance issues or timeouts
-- Build or compilation errors
-
-## Debugging Protocol
-
-### 1. Reproduce the Issue
-
-Before fixing, ensure you can reproduce:
-
-```bash
-# Run the failing command/test again
-# Note exact error output
-# Capture any stack traces
-```
-
-**Record what you find** - Use `swarm_accumulate_error` to track errors for retry context.
-
-### 2. Gather Context
-
-Read surrounding code and error context:
-
-- Error message and stack trace
-- Recent changes to affected files
-- Related configuration files
-- Test fixtures and setup
-
-### 3. Form Hypotheses
-
-Based on the error, list possible causes:
-
-1. Most likely cause based on error message
-2. Recent changes that could cause this
-3. Environmental/configuration issues
-4. Edge cases or missing validations
-
-### 4. Test Hypotheses Systematically
-
-For each hypothesis:
-
-1. Add targeted logging or assertions
-2. Modify one variable at a time
-3. Verify if behavior changes
-4. Document findings
-
-### 5. Implement Fix
-
-Once root cause is identified:
-
-1. Write the minimal fix
-2. Add regression test
-3. Verify original error is gone
-4. Check for similar issues elsewhere
-
-### 6. Document Learning
-
-Use `semantic-memory_store` if you discovered a pattern worth preserving for future sessions.
-
-## Common Error Patterns
-
-### Type Errors
-
-```
-Error: Cannot read property 'x' of undefined
-```
-
-**Check**: Variable initialization, null/undefined checks, async timing
-
-### Import Errors
-
-```
-Error: Module not found
-```
-
-**Check**: Path correctness, package installation, export statements
-
-### Timeout Errors
-
-```
-Error: Timeout of 5000ms exceeded
-```
-
-**Check**: Async operations, network calls, infinite loops
-
-### Validation Errors
-
-```
-Error: Validation failed for field 'x'
-```
-
-**Check**: Input data, schema definitions, required fields
-
-## Debugging Tools
-
-### Console Logging
-
-```typescript
-console.log("[DEBUG]", { variable, context });
-```
-
-### Breakpoints (if interactive)
-
-- Add `debugger;` statement
-- Use IDE breakpoints
-
-### Binary Search
-
-For "it was working before":
-
-1. Find last known good commit
-2. Binary search through commits
-3. Identify the breaking change
-
-## Swarm Integration
-
-When debugging in a swarm:
-
-1. **Report blocker immediately** - Don't spin alone
-2. **Share context** - Use `swarm_broadcast` with findings
-3. **Create beads** - Track discovered issues as bugs
-4. **Check other agents** - They might have relevant context
-
-## Anti-Patterns
-
-- **Shotgun debugging** - Random changes hoping something works
-- **Silent fixes** - Fixing without understanding root cause
-- **Ignoring warnings** - Warnings often predict errors
-- **Assuming environment** - Works on my machine != works everywhere