134 skills · sorted by relevance
Review MSBuild project files and build definitions for correctness and maintainability. Get prioritized findings with concrete recommendations, distinguishing confirmed defects from conditional risks and valid exceptions such as F# source ordering or required imports.
Walk Chinese-speaking users through code or a pull request while checking correctness. Follow the actual execution flow, explain key implementation choices, and report concrete defects with trigger conditions, consequences, and precise file locations.
Review a specified Marin pull request for newly introduced defects, repository instruction violations, and concrete maintenance problems. Receive findings backed by code evidence, with optional posting of review comments to GitHub.
Reproduce and fix LazyCodex or Codex defects in a temporary workspace. Verify the change with regression tests and real usage, then deliver an upstream Codex pull request or a LazyCodex issue containing the tested patch.
Qualify a check before it is allowed to clear anything — prove it can detect the failure it is meant to catch. Seeds known defects into a copy of a real artifact plus a clean control, runs the checker, and reports recall and false-positive rate into a qualification ledger. Use when the user says "does this check work", "qualify the gate", "test my reviewer", "seed defects", "vaccinate", "qualify the checks", "can I trust this review", "does it pass for the right reason", or before relying on any automated check or referee simulation for a decision that matters. NOT a code fixer and NOT a reviewer itself — it grades the grader.
The English summary is being prepared. Showing the original description.
Comprehensive adversarial audit of a theory, proof, math/econ paper, codebase, or set of claims — decompose into components, fan out independent skeptics that must return CONCRETE defects, adjudicate every finding with a separate judge, fix all confirmed defects, then re-verify. Use when correctness must be bulletproof and single-pass or round-by-round review is too slow and too shallow. Invoke for "audit this rigorously", "find ALL the bugs/gaps", "make this rock solid", "converge faster on correctness".
The English summary is being prepared. Showing the original description.
Hunts bugs with evidence trails. Use when investigating unexpected behavior or before merging code with potential hidden defects.
The English summary is being prepared. Showing the original description.
Reproduce and repair a bounded Proto UI regression whose expected behavior and owning layer are already governed. Use this before domain implementation skills whenever work starts from a reproducible failure, including bug fixes, drift repairs, broken exports, Runtime failures, Adapter parity defects, Prototype defects, or public-projection regressions. Do not use when expected semantics or ownership is unresolved.
The English summary is being prepared. Showing the original description.
Triage defects with independent severity and priority, reproducible evidence, preserved attempt history, and explicit resolution retesting.
The English summary is being prepared. Showing the original description.
Coordinate repairs across multiple OpenClaw issues or pull requests, which propose code changes. Assign owners, prove defects, verify fixes, and track outcomes through merging or closing resolved items within the authorized scope.
Investigate OpenClaw issues and review pull requests, proposed code changes, using current source and test evidence. Depending on the authorized scope, repair defects locally or carry the change through review, merge, and final verification.
Independently reviews deliverables handed over through Kanban, checks them against acceptance criteria and test evidence, and records an approval, actionable change request, or escalation when a human decision or external prerequisite is needed.
Review a design across five areas: visual execution, fit to the user's problem, brand consistency, accessibility, and wording. It guides the agent to fix concrete defects and explain any issues that remain unresolved.
Organize a GitHub issue backlog by shared root cause, with one main tracking issue and design document per group. Route new reports into existing plans and prepare pull requests that address related bugs together with regression checks.
Evaluate whether tests catch defects in business logic, calculations, or state machines. Using Stryker, the skill introduces small code changes, reviews which changes escape detection, and helps strengthen tests where important behavior is insufficiently checked.
Review form validation so users, including people using screen readers, can understand and correct errors. Check field descriptions, required states, and grouped choices in rendered HTML, while accounting for React forms that handle submission in the browser.
Review pull requests, branches, commit ranges, or local changes with GitNexus. Combine code relationship analysis with source and test inspection to identify concrete defects, missing tests, affected behavior, and risks before merging.
Document bug fixes and incident recovery with a structured receipt covering the problem, cause, change, and verification evidence. Distinguishes verified results from partial or blocked work, identifies missing proof, and specifies the next check needed to close the gap.
Review logs and source files from a completed manual example run. Check that results are current and complete, verify intended behavior beyond successful exit codes, and distinguish product defects from environment or credential problems.
Re-review a pull request after new commits arrive. Check every previous finding, inspect the new changes for regressions, and review the full current change. Run relevant verification and publish a test report and review verdict within Factory.
Review open Mastra GitHub issues and trace reported defects to source code. Apply the @mastra/core label to confirmed core bugs, report uncertain ownership, and summarize whether fixes are active, stale, or unclaimed.
Find unused components, files, translation keys, and other unnecessary code across a codebase. Document every consumer and removal impact in a tracking issue, then turn one maintainer-approved candidate into a focused pull request.
Review agent-written code changes before merge, focusing on concrete correctness defects. Uses independent subagent reviews, targeted verification, and fresh checks after substantial repairs, with a deeper evidence-based process for security, concurrency, persistence, or data-loss risks.
Review code changes or a repository for actionable bugs by tracing how components interact. Findings explain reachable triggers, incorrect behavior, consequences, and checked counterarguments. Correctness is the default focus; performance and security can also be selected.