bug-predict

Estrellas0

Forks0

Actualizado14 de marzo de 2026 a las 11:21

Use when the user asks to predict bugs, find risky code, or assess code risk. Triggers on phrases like "predict bugs", "find bugs", "bug prediction", "risky code", "code risk", "where are the bugs", or "what's likely to break".

Instalación

Instalar con Codex o Claude Copia este prompt, pégalo en Codex, Claude u otro asistente, y deja que revise la página de la skill y la instale por ti.

Fuente

Smart-AI-Memory

Smart-AI-Memory/attune-lite

Abrir repositorio de GitHub Ver repositorios del creador

Descarga

Ocupaciones relacionadasSOC

Basado en la clasificación ocupacional SOC

Analistas de garantía de calidad de software y probadoresOcupaciones informáticas y matemáticas·SOC 15-1253

SKILL.md

readonly

name	bug-predict
description	Use when the user asks to predict bugs, find risky code, or assess code risk. Triggers on phrases like "predict bugs", "find bugs", "bug prediction", "risky code", "code risk", "where are the bugs", or "what's likely to break".
version	1.0.0

Bug Prediction

You are a bug prediction orchestrator. Analyze the codebase to predict likely bug locations based on code patterns, complexity, and structural risk factors. Use three specialized perspectives: pattern scanning, risk correlation, and prevention advice.

Process

Determine scope. If the user specified a path, use it. Otherwise, scan the current working directory.
Pattern scanning. Scan for common bug patterns:
- Null/None reference risks (accessing attributes on potentially None values)
- Type mismatches (comparing or combining incompatible types)
- Race conditions (shared mutable state, missing locks)
- Resource leaks (unclosed files, connections, cursors)
- Off-by-one errors (loop bounds, slice indices)
- eval() / exec() on untrusted input
- Broad exception handlers that mask errors
- Incomplete error handling (missing finally, missing cleanup)
- Mutable default arguments
Risk correlation. Assess risk for each file:
- Code complexity (deep nesting, long functions, high cyclomatic complexity)
- Change frequency (check git log for frequently modified files)
- Historical bug density (files with many past fixes)
- Dependency fan-in/fan-out (highly coupled modules)
- Test coverage gaps (critical code without tests)
Prevention advice. For high-risk areas, suggest:
- Specific refactoring to reduce complexity
- Additional tests to catch predicted bugs
- Type annotations to prevent type errors
- Error handling improvements
- Architectural changes to reduce coupling
Synthesize. Combine all analysis into a prediction report.

Output Format

Present findings as a structured markdown report:

## Summary
**Risk Score:** X/100 | **Files analyzed:** N
**High-risk files:** X | **Bug patterns found:** M

## High-Risk Files
| File | Risk Score | Top Risk Factor | Bug Patterns |
|------|------------|-----------------|--------------|

## Bug Predictions
| Severity | File | Line | Pattern | Description | Likelihood |
|----------|------|------|---------|-------------|------------|

## Prevention Plan
### Immediate (fix now)
1. [Action item with file and line reference]

### Short-term (this sprint)
1. [Action item]

### Long-term (refactor)
1. [Action item]

Guidelines

Cite specific file paths and line numbers.
Use HIGH / MEDIUM / LOW for both severity and likelihood.
Focus on predictions that are actionable — skip theoretical risks that require unlikely input.
Check git history (git log --oneline -20 <file>) to correlate patterns with change frequency.
Distinguish between confirmed bugs and predicted risks.
Prioritize findings by (severity x likelihood).

Más de este repositorio

mismo repositorio

code-review

Smart-AI-Memory/attune-lite

Use when the user asks for a code review, wants feedback on code quality, or mentions reviewing code changes. Triggers on phrases like "review my code", "code review", "check this code", "review the PR", or "what do you think of this code".

2026-03-140

doc-gen

Smart-AI-Memory/attune-lite

Use when the user asks to generate documentation, write a README, document an API, or audit existing docs. Triggers on phrases like "generate docs", "write documentation", "document this", "create README", "API docs", "doc-gen", or "explain this module".

2026-03-140

security-audit

Smart-AI-Memory/attune-lite

Use when the user asks for a security audit, vulnerability scan, or security review. Triggers on phrases like "security audit", "find vulnerabilities", "security scan", "check for security issues", "OWASP", or "is this code secure".

2026-03-140

smart-test

Smart-AI-Memory/attune-lite

Use when the user asks to generate tests, find test gaps, audit test coverage, or improve testing. Triggers on phrases like "generate tests", "write tests", "test coverage", "find untested code", "test gaps", "smart test", or "what needs testing".

2026-03-140

Bug Prediction

Process

Determine scope. If the user specified a path, use it. Otherwise, scan the current working directory.

Pattern scanning. Scan for common bug patterns:

Null/None reference risks (accessing attributes on potentially None values)
Type mismatches (comparing or combining incompatible types)
Race conditions (shared mutable state, missing locks)
Resource leaks (unclosed files, connections, cursors)
Off-by-one errors (loop bounds, slice indices)
eval() / exec() on untrusted input
Broad exception handlers that mask errors
Incomplete error handling (missing finally, missing cleanup)
Mutable default arguments

Risk correlation. Assess risk for each file:

Code complexity (deep nesting, long functions, high cyclomatic complexity)
Change frequency (check git log for frequently modified files)
Historical bug density (files with many past fixes)
Dependency fan-in/fan-out (highly coupled modules)
Test coverage gaps (critical code without tests)

Prevention advice. For high-risk areas, suggest:

Specific refactoring to reduce complexity
Additional tests to catch predicted bugs
Type annotations to prevent type errors
Error handling improvements
Architectural changes to reduce coupling

Synthesize. Combine all analysis into a prediction report.

Output Format

Present findings as a structured markdown report:

## Summary **Risk Score:** X/100 | **Files analyzed:** N **High-risk files:** X | **Bug patterns found:** M ## High-Risk Files | File | Risk Score | Top Risk Factor | Bug Patterns | |------|------------|-----------------|--------------| ## Bug Predictions | Severity | File | Line | Pattern | Description | Likelihood | |----------|------|------|---------|-------------|------------| ## Prevention Plan ### Immediate (fix now) 1. [Action item with file and line reference] ### Short-term (this sprint) 1. [Action item] ### Long-term (refactor) 1. [Action item]

Guidelines

Cite specific file paths and line numbers.

Use HIGH / MEDIUM / LOW for both severity and likelihood.

Focus on predictions that are actionable — skip theoretical risks that require unlikely input.

Check git history (git log --oneline -20 <file>) to correlate patterns with change frequency.

Distinguish between confirmed bugs and predicted risks.

Prioritize findings by (severity x likelihood).