More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    automated gameplay qa bug reproduction agent

    1

    Transform fragmented gameplay bug evidence into professional reproduction procedures and regression test suites.

    $10.99

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    automated gameplay qa bug reproduction agent

    automated gameplay qa bug reproduction agent

    Example session with this skill installed

    Convert this report: Player says the game froze during the boss transition in build 1.2. Log shows 'Warning: missing phase target' at 12:14:16. They were playing on Windows with a gamepad.

    • Read your context and instructions
    • Compiled the automated gameplay qa

    The provided evidence has been normalized into a structured defect package. The failure is classified as a Soft Lock rather than a crash, as the process remained active but progression halted.

    [Quest] Phase transition during Warden boss encounter results in Soft Lock

    Build: 1.2
    Platform: Windows
    Severity: S1 (Blocker)
    Frequency: Unknown (Single report)

    Reproduction Steps

    1. Launch Build 1.2 on Windows.
    2. Load a save file at the 'Warden Arena' entrance.
    3. Engage the Warden boss and reduce health to 50%.
    4. Observe the transition to Phase 2.
    5. Attempt to trigger the transition while providing gamepad input.

    Expected Behavior

    The boss should play a transition animation, its health bar should refresh, and the arena exit should remain locked until the boss is defeated.

    Actual Behavior

    The boss entity is removed from the scene. A log warning Warning: missing phase target is generated at 12:14:16. The arena exit remains locked, but no new boss entity spawns, resulting in a progression blocker (Soft Lock).

    Suspected System

    Primary: Quest / Mission Scripting (High Confidence)
    Secondary: Boss State Machine / Spawner (Medium Confidence)

    Evidence Summary

    Log Evidence: 12:14:16.188 — Warning: missing phase target indicates a failure in the script responsible for spawning the Phase 2 actor.

    • User Report: Confirms the game remained running (HUD visible) but interaction with the boss was lost.

    Regression Tests

    • Test-01: Verify boss phase transition completes without gamepad input.
    • Test-02: Verify transition with varied frame caps (30/60/120 FPS) to check for timing races.
    • Test-03: Verify save/load persistence after the phase transition event.

    Next steps

    • Provide a full engine log to check for preceding memory or asset-loading warnings.
    • Attempt reproduction 10 times to establish a frequency baseline.
    • Test if the inventory UI was opened during the transition, as this often interrupts state scripts.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Bug reports in game development are often vague, missing critical environment data, or failing to capture the exact timing of a state-driven failure. This skill transforms fragmented evidence into professional QA defect packages that engineering can actually use to fix the problem.

    What it does

    • Analyze evidence by processing logs, crash reports, save files, and gameplay recordings to extract technical facts.
    • Generate reproduction steps that are chronological, minimal, and executable by other testers or developers.
    • Categorize defects into specific severity levels (S0 to S5) based on player impact rather than implementation difficulty.
    • Design test matrices for intermittent, multiplayer, or performance-related bugs to isolate variables like frame rate or latency.
    • Draft regression tests and verification checklists to ensure fixes are robust and do not introduce new regressions.

    How it works

    1. Intake evidence by providing the agent with raw logs, screenshots, build metadata, or rough player reports.
    2. Normalize facts as the agent separates direct observations from assumptions and identifies missing environment data.
    3. Execute reproduction analysis to determine the minimum state needed to trigger the bug, including save versions and quest flags.
    4. Produce the QA package containing structured steps, expected vs. actual behavior, and suspected subsystems.

    Frameworks & tools

    This tool is engine-agnostic and supports Unity, Unreal Engine, Godot, and proprietary engines. It integrates with workflows involving Jira, GitHub Issues, crash reporting tools, and performance profiling captures.

    Why this beats prompting it yourself

    Generic LLMs often hallucinate missing steps or fail to account for game-specific state dependencies like frame-time spikes and network authority. This skill uses a specialized taxonomy for gameplay failures, ensuring that desyncs, soft locks, and save migrations are handled with technical precision.

    Use cases

    • Progression blockers where an NPC disappears or a quest trigger fails during specific transitions.
    • Multiplayer desync between host and client roles under high latency or packet loss conditions.
    • Save corruption analysis across different build versions and cloud-sync conflicts.
    • Performance-related failures where frame rate hitching causes physics or input instability.

    Known limitations

    Cannot guarantee reproduction without access to the running build. Relies on the quality of provided logs and timestamps. Does not replace studio-specific risk-acceptance policies.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Testing & Debugging.

    Requires runtime evidence (logs, metadata) for high confidence. Cannot execute builds. Studio-specific priority rules and risk-acceptance require manual adjustment.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    Accounts
    Google DriveLinearGitHubNotionSlack

    Permissions

    Read Files
    Browser
    Write Files

    Allowed Hosts

    drive.googleapis.com
    api.linear.app
    api.github.com
    api.notion.com
    api.slack.com

    File Scopes

    **/*.md
    **/*.txt
    **/*.json
    **/*.jsonl
    **/*.yaml
    **/*.yml
    **/*.xml
    **/*.ini
    **/*.cfg
    **/*.config
    **/*.csv
    **/*.tsv
    **/*.html
    **/*.log
    **/*.dmp
    **/*.dump
    **/*.crash
    **/*.trace
    **/*.utrace
    **/*.png
    **/*.jpg
    **/*.jpeg
    **/*.webp
    **/*.gif
    **/*.mp4
    **/*.mov
    **/*.avi
    **/*.mkv
    **/*.webm
    **/*.sav
    **/*.save
    **/*.bin
    **/*.dat
    **/*.db
    **/*.sqlite
    **/*.pdf
    **/*.doc
    **/*.docx
    **/*.xls
    **/*.xlsx
    **/qa/**
    **/quality-assurance/**
    **/bugs/**
    **/bug-reports/**
    **/issues/**
    **/triage/**
    **/reproduction/**
    **/repro/**
    **/regression/**
    **/regression-tests/**
    **/verification/**
    **/test-cases/**
    **/test-plans/**
    **/test-results/**
    **/playtests/**
    **/playtest/**
    **/gameplay/**
    **/recordings/**
    **/videos/**
    **/screenshots/**
    **/captures/**
    **/logs/**
    **/crashes/**
    **/crash-reports/**
    **/crash-dumps/**
    **/stack-traces/**
    **/telemetry/**
    **/performance/**
    **/profiling/**
    **/memory/**
    **/fps/**
    **/frame-time/**
    **/saves/**
    **/save-files/**
    **/save-system/**
    **/checkpoints/**
    **/progression/**
    **/quests/**
    **/missions/**
    **/inventory/**
    **/combat/**
    **/abilities/**
    **/physics/**
    **/collision/**
    **/animation/**
    **/ai/**
    **/navigation/**
    **/ui/**
    **/hud/**
    **/input/**
    **/multiplayer/**
    **/network/**
    **/networking/**
    **/replication/**
    **/matchmaking/**
    **/servers/**
    **/backend/**
    **/builds/**
    **/build-info/**
    **/releases/**
    **/branches/**
    **/commits/**
    **/ci/**
    **/automation/**
    **/automated-tests/**
    **/test-harness/**
    **/unity/**
    **/unreal/**
    **/godot/**
    **/windows/**
    **/macos/**
    **/linux/**
    **/android/**
    **/ios/**
    **/console/**
    **/localization/**
    **/streaming/**
    **/assets/**
    **/entitlements/**
    **/dlc/**
    **/docs/**
    **/documentation/**
    **/output/**
    Listed1 month ago

    What's inside

    Frequently Asked Questions

    This agent is engine-agnostic and works with Unity, Unreal Engine, Godot, and proprietary setups, as long as you can provide standard text-based logs, crash reports, or build metadata.

    Unlike generic LLMs, this agent uses a specialized gameplay failure taxonomy to account for state-driven issues like network authority, frame-time spikes, and save migrations without hallucinating missing steps.

    You receive the specialized QA agent configuration, a suite of structured templates for different bug categories (e.g., physics, networking, UI), and a guide on optimizing your log exports for best results.

    No complex installation is required; simply integrate the agent into your existing LLM workspace and begin uploading your raw evidence files or report drafts for processing.

    The agent can analyze gameplay recordings and logs to identify timing and state triggers, though it cannot physically execute the game; it provides the precise roadmap for your team to replicate the issue.

    New in Testing & Debugging