Test design, vulnerability assessment, and quality assurance (13K skills)
Review software before release by checking features, documentation, tests, code quality, and security. Use the Centinela QA agent to produce a versioned assessment with a release decision and specific reasons for any blockers.
Review a codebase for unused code, outdated or vulnerable dependencies, and patterns that make maintenance harder. Produce a prioritized Markdown report, compare recurring findings with earlier scans, and prepare a handoff for the development agent.
Review code for security issues in authentication, access controls, data handling, embedded secrets, and dependencies. The Centinela QA agent produces a report with severity ratings, an approval verdict, and recommended repair priorities.
Review and strengthen Go code against injection attacks, unsafe file access, authentication flaws, and secret exposure. Traces how untrusted data reaches sensitive operations, ranks findings by severity, and guides fixes, static analysis, and security testing.
Write, review, audit, and debug Go tests that check observable behavior. Covers named test cases, integration test isolation, concurrent execution, race and goroutine leak detection, and finding causes of intermittent failures or gaps in test coverage.
Set up golangci-lint to check Go code for potential problems. Helps select checks, configure project rules, understand warnings, apply available automatic fixes, and document justified exceptions when a warning needs to be suppressed.
Write and review Go tests using stretchr/testify. Choose assertions that continue or stop on failure, create mocks that verify expected calls, and organize related tests with shared setup and cleanup routines.
Check whether an implementation or generated output meets its stated goal. Review files, required content, logical consistency, and project conventions, then receive PASS, WARN, or FAIL results with specific issues to fix.
Review a coding agent’s work against its assigned requirements, permitted file changes, and verification evidence. Return a structured verdict with supporting reasons, identifying incomplete work, scope violations, or missing proof without modifying the reviewed files.
Evaluate a CCB execution round using task requirements, coordinator summaries, worker and checker reports, and command or test evidence. Produce a machine-readable decision distinguishing success, limited repair, partial completion, replanning, and blockers requiring intervention.
Review one task result returned by a CCB worker agent against its assigned scope, acceptance criteria, and expected verification. Identify hidden fallback behavior, reduced quality or scope, and missing evidence, then report the outcome and required rework.
Assess a completed execution round from supplied review, integration, change promotion, and verification records. Check for missing or conflicting evidence and return a script-readable result with audit details, without running tools or modifying code.
Find reference solutions for CTF security challenges using challenge names, distinctive code, or error messages. Search for related writeups, extract promising pages, and compare their techniques to decide which approach may fit the current challenge.
Start unfamiliar CTF challenges by examining files, descriptions, and remote services, then route the work to an appropriate specialist skill. Guides category changes when progress stalls, checks candidate answers, and hands off to a solution write-up skill.
Investigate AI and machine learning CTF challenges by inspecting model weights, comparing training changes, and exploring adversarial inputs that cause misclassification. Includes references for prompt injection, model reconstruction, and identifying possible training data.
Analyze CTF evidence such as disk images, memory dumps, logs, and network captures to recover files and trace activity. Includes techniques for finding hidden data in images, audio, and PDFs, plus hardware signal analysis.
Investigate CTF clues through OSINT, research using publicly available information. Use image searches, social profiles, DNS records, and archived pages to identify locations, trace accounts, and connect evidence from multiple public sources.
Develop exploits for CTF binaries with known memory handling weaknesses. Inspect protections, debug failures, and choose techniques for stack, heap, or kernel bugs to turn an identified flaw into code execution or elevated privileges.
Understand how CTF challenge programs work through reverse engineering, the analysis of compiled software. Use guidance on binaries, APKs, WASM, firmware, and custom virtual machines to trace validation logic, investigate obfuscation, and choose analysis tools.
Explore CTF puzzles that span categories or use unusual formats, including layered encodings, restricted Python or Bash environments, QR codes, audio signals, and logic games. Provides technique references for decoding clues and solving constraints.
Work through CTF challenges involving web applications, APIs, and browser behavior. Map requests and hidden features, investigate injection or authentication flaws, and build small proofs before combining weaknesses into a path to the challenge flag.
Analyze malware in CTF security challenges using techniques for obfuscated scripts, executable files, encrypted traffic, and memory dumps. Follow references to extract configurations, identify command-and-control servers, and find indicators of compromise, such as suspicious domains.
Analyze cryptography challenges in CTF security competitions by matching clues to known weaknesses in encryption, signatures, and random number generators. Use technique summaries and supporting code examples to work toward recovering plaintext or keys.
Use an unfinished teaching template to practice test-first development, building one behavior at a time. It defines how to verify a failing test, while leaving implementation and refactoring steps for the learner to complete.