pre-push-review_skill

Overview

This skill performs a focused pre-push review of all local commits that have not yet been pushed to the remote branch. It inspects commit differences, checks implementation against available specifications, runs automated validators, and produces actionable, prioritized feedback before code leaves the developer machine. The goal is to catch correctness, quality, testing, and documentation gaps early and provide a clear readiness verdict.

How this skill works

It first identifies unpushed commits by comparing the current branch to its remote tracking branch and lists the commits to be reviewed. It searches for a matching feature specification and, if found, verifies the implementation against requirements, tasks, and design notes. The skill runs linters, formatters, and project validation commands, inspects tests and docs, and generates an implementation explanation and completeness assessment when a spec exists. Finally, it produces categorized findings (Critical, Important, Minor, Suggestion) with concrete remediation steps and a push recommendation.

When to use it

• Before pushing any feature branch with new or modified commits
• When finishing a bugfix branch to ensure tests and error handling are adequate
• Prior to an integration window or CI run to reduce noise from avoidable failures
• When a spec or formal requirement exists and you need a traceable implementation check
• If the change introduces new public APIs, tooling changes, or infra config updates

Best practices

• Run the skill locally as the final pre-push step to catch regressions early
• Ensure branch names or a specs directory include clear feature identifiers to enable spec matching
• Keep tests focused on behavior, add coverage for edge cases introduced by the change
• Document any intentional divergence from the spec with rationale in project decision logs or code comments
• Keep linters and project validation commands in the repo so automated checks are reproducible

Example use cases

• A developer finished a feature and wants verification that all spec tasks are implemented and tested
• A bugfix branch that modifies core logic needs a review for edge cases and error handling before push
• A contributor adds a new CLI command or config and needs documentation and API surface checks
• A team member is unsure whether unit tests adequately cover newly introduced behavior and wants a coverage and test-quality assessment