Exécutez n'importe quel Skill dans Manus
en un clic

Exécutez n'importe quel Skill dans Manus en un clic

sentinel-review

Étoiles0

Forks0

Mis à jour29 avril 2026 à 07:00

Structured code review for the current diff or specified files. Flags bugs, logic errors, security issues, and style violations with severity levels. Use before committing, before PRs, or when reviewing changes.

Installation

Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.

Exécuter dans Manus

Source

89jobrien

89jobrien/warpx

Ouvrir le dépôt GitHub Voir les dépôts du créateur

Téléchargement

Exécuter dans Manus

Métiers associésSOC

Basé sur la classification professionnelle SOC

Analystes en assurance qualité des logiciels et testeursProfessions informatiques et mathématiques·SOC 15-1253

SKILL.md

readonly

name	sentinel-review
description	Structured code review for the current diff or specified files. Flags bugs, logic errors, security issues, and style violations with severity levels. Use before committing, before PRs, or when reviewing changes.

Sentinel Review

Read-only code reviewer. Flags issues, does not fix them.

Process

Identify the diff to review (staged changes, branch diff, or specified files)
Classify each finding: blocking, suggestion, or nitpick
Report all findings in a single pass
Include file path, line number, severity, and explanation for each finding

Review Categories

Correctness: logic errors, off-by-one, null/None handling
Security: injection, secrets in code, unsafe patterns
Performance: unnecessary allocations, O(n^2) where O(n) suffices
Style: naming, dead code, missing error context
Architecture: coupling, wrong abstraction layer, missing tests

Rules

Apply ALL severity levels in one pass before reporting
Do not propose fixes inline -- just flag and explain
False positives in test files (variable named password, localhost IPs) should be noted but not flagged as blocking

Plus depuis ce dépôt

même dépôt

writing-plans

89jobrien/warpx

Write an implementation plan as a markdown file, then immediately serialize it to a timestamped YAML task file in .ctx/tasks/. Use before touching any code when given a spec or approved design.

2026-05-030

godmode-brainstorm

89jobrien/warpx

Design-first workflow for new features, components, or architecture decisions. Use before writing any code — get explicit approval on a design first.

2026-05-030

godmode-cap

89jobrien/warpx

Commit-and-push workflow with validation. Use when asked to "cap", "commit and push", or "ship it".

2026-05-030

godmode-debug

89jobrien/warpx

Systematic debugging for Rust. Use before proposing any fix — establish root cause first, then write a failing test, then fix.

2026-05-030

godmode-parallel

89jobrien/warpx

Dispatch parallel subagents for independent tasks. Use when 2+ tasks have no shared state or sequential dependencies.

2026-05-030

godmode-tasks

89jobrien/warpx

Task graph management for a session. Use to create tasks, track progress, execute the next unblocked task, or manage dependencies between work items.

2026-05-030

name	sentinel-review
description	Structured code review for the current diff or specified files. Flags bugs, logic errors, security issues, and style violations with severity levels. Use before committing, before PRs, or when reviewing changes.

Sentinel Review

Read-only code reviewer. Flags issues, does not fix them.

Process

Identify the diff to review (staged changes, branch diff, or specified files)
Classify each finding: blocking, suggestion, or nitpick
Report all findings in a single pass
Include file path, line number, severity, and explanation for each finding

Review Categories

Correctness: logic errors, off-by-one, null/None handling
Security: injection, secrets in code, unsafe patterns
Performance: unnecessary allocations, O(n^2) where O(n) suffices
Style: naming, dead code, missing error context
Architecture: coupling, wrong abstraction layer, missing tests

Rules

Apply ALL severity levels in one pass before reporting
Do not propose fixes inline -- just flag and explain
False positives in test files (variable named password, localhost IPs) should be noted but not flagged as blocking