Test design, vulnerability assessment, and quality assurance (13K skills)
Run an audit, separate findings that can be fixed automatically from those needing manual work, and apply eligible fixes with tests and separate commits. Use severity filters, a finding limit, or a classification-only dry run.
Audit a completed AI development phase to check whether its planned evaluations were implemented. Produce EVAL-REVIEW.md with a score, verdict, identified gaps, and a plan to address missing coverage.
Add tests to a completed development phase using its records and implementation. Review the proposed coverage, then generate unit tests for individual components and browser tests for full workflows, verify them, and commit the test files.
Audit acceptance tests, which check whether features meet user needs, and verification records across project phases. Compare outstanding items with the codebase, flag outdated documentation, and produce a prioritized plan for human testing.
Routes quality checks to the appropriate skill, covering code review and fixes, debugging, security checks, acceptance testing, AI response evaluation, and interface reviews. It also directs requests for validating phase outputs or reviewing plans across AI systems.
Validate Claude Ads data contracts, scoring inputs, run outputs, installation procedures, and release readiness. Return machine-readable pass or fail results with evidence, the most urgent blocker, and recovery steps, while checking source freshness and safe removal boundaries.
Model distributed systems and concurrent algorithms in TLA+ or PlusCal, then check possible states with TLC. Use the results to investigate deadlocks, broken rules, and differences between specifications and implementation code.
Write Apache Cassandra distributed tests that simulate multiple nodes within one JVM, the Java runtime. The guide covers cluster setup, queries, node restarts, message dropping for failure scenarios, and troubleshooting problems caused by isolated class loading.
Turn bug reports into small, runnable tests or scripts with clear failure checks. Verify that they expose the reported problem, then reduce inputs and setup while preserving the same failure.
Write and review Apache SeaTunnel unit tests and E2E tests that check complete jobs using Testcontainers. It provides concrete rules for container setup, waiting, resource cleanup, and assertions to address intermittent failures and resource leaks.
Audit npm dependencies before updating them with Bun. Check maintainers, release age, signatures, published file changes, and release notes, then update packages that pass review and record deferred updates for later follow-up.
Review Renovate pull requests that update GitHub Actions. Verify pinned commits against upstream version tags, inspect release notes for breaking changes, and check existing workflow settings before giving a clear recommendation on whether to merge.
Record real ClawHub browser behavior and turn it into evidence for a GitHub PR. Add captions and focused zooms, distinguish before and after results, inspect the finished media, and attach it directly to the PR.
Investigate Convex production errors, prepare fixes on a branch, and verify them before opening a pull request for human review. After an approved deployment, check whether the same error recurs. Never automatically merge changes or deploy to production.
Tests a specific Convex feature with sample data and simulated users, without deployment. It checks expected results, rejects unauthorized callers, and verifies that queries return only data belonging to the current user.
Combine Convex authorization, code quality, performance, and failure audits into one readiness report. It removes duplicate findings, scores confirmed issues with a stated formula, and orders the fixes while clearly identifying checks that could not run.
Review code in a convex/ directory for security, access control, input and output validation, performance, and schema design. Receive findings grouped by severity, with explanations of each issue and suggested fixes before release.
Create tests for Convex functions with convex-test and vitest, using a backend that runs in memory. Check arguments, return values, authentication, indexes, errors, and scheduled functions without relying on real time or network access.
Audit and fix four access-control problems in Convex apps: impersonation, missing ownership checks, sensitive data exposure, and writes into another user's container. Check the authentication foundation first, then verify changes with TypeScript.
Design or extend DataFusion benchmarks using SQL where practical. Select representative workload variations, keep unrelated processing lightweight, and integrate runs with bench.sh so measurements reflect useful execution paths and remain practical to run.
Audit one datafusion-spark function against Apache Spark 4.2.0. Compare source code, accepted types, and edge cases, verify expected results with PySpark, then fix differences and record remaining gaps as issues and disabled tests.
Review codebases, APIs, and tools for vulnerabilities using source evidence and bounded local checks. Produces supported findings with impact, priority, and focused fixes, and can create structured reports for explicitly requested full audits.
Compare Bifrost HTTP API code with OpenAPI, its API specification, to find missing routes and mismatched methods, authentication, parameters, and responses. Get an audit report with code references and recommended fixes.
Turn a Bifrost PR or GitHub issue into API regression tests that catch recurring bugs. Review existing coverage, add cases to the Postman collection, and validate collection processing and provider filtering.