classic-game-design-test-8aa7ff — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited classic-game-design-test-8aa7ff (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
This skill protects the user's actual intent: "I want to play the old game I remember" is not the same as "make a remake." A remake, remix, clone, or browser-inspired fallback can be useful, but it must be labeled as such and must not be counted as satisfying a request for the real childhood game.
The target experience is a chatbot/browser user saying the game they remember, the workflow finding the rights-cleared original or best compatible public release, adapting it to the current machine without manual emulator setup, and ending with a single launch surface that plays immediately.
compatibility patch, remake, or fallback.
gameplay.
prevent the real game path.
frontend-ui-engineering.browser-testing-with-devtools.test-driven-development.ui-test when the chatbot surface changes.documentation-and-adrs.Before building, classify the requested outcome:
runtime, emulator, official web player, or source port.
modern engine or wrapper.
externally acquired.
If the user asked for the real old game, pursue Original game path first. Only switch to another class when the blocker is explicit and documented.
Do not use warez, cracked images, copyrighted firmware, or private downloads. For every game/runtime/media source, record:
If the only available route requires copyrighted firmware or game media the project cannot provide, mark the task NOT_PLAYABLE_YET instead of pretending a remake satisfies it.
screenshots, online references, and titles with similar names. Record the confidence and alternatives.
Keep remakes/fallbacks subordinate unless the user explicitly requested one.
compatibility layer, or library for the core platform logic. Do not hand-roll CPU, disk, audio, physics, parsing, or AI engines when a proven runtime exists.
user to install a desktop app, restart, manually boot an emulator, hunt for firmware, or copy files unless that limitation is explicitly accepted.
The launcher must wait for the runtime to be ready, inject media once, avoid reload loops, and expose status that distinguishes "loaded" from "playing." If the runtime stages media in a file slot, use its explicit boot/mount with reset semantics; do not count a post-boot insert as an autostart proof. For emulator APIs that acknowledge disk/ROM insertion asynchronously, wire a mount acknowledgement or bounded retry. Never schedule blind repeated resets; a reset loop is a blocker even when the disk eventually appears mounted.
mappings needed by the target game. Include a real browser click/tap path for any action the user naturally expects to perform with the mouse.
Provide an explicit launch/play gesture, then verify the audio context is running and unmuted. Do not claim sound works from code inspection alone.
emulator ready, game started, input accepted, audio running, and fatal blockers. Avoid mock success states that look playable.
it as a fallback/remix and keep it separate from the real-game acceptance result.
"Playable" means a real gameplay objective was achieved in the live runtime. For an artillery/tank game, the minimum proof is:
or explosion.
For other genres, replace "hit a tank" with the smallest genre-authentic proof: move a character and collect an item, complete a menu-start-to-gameplay loop, win/lose a round, enter a level, or trigger a visible scored action.
Never claim final playability from:
Observation-only test instrumentation is allowed, but it must not change game state to manufacture the proof. The user-facing input path, such as a real button or canvas click, still has to trigger the gameplay action.
Use browser-testing-with-devtools or an equivalent real browser path. Capture evidence with date, environment, URL, and exact build/version.
mounted disk is not proof that Startup-Sequence ran.
AudioContext running/unmuted, andobservable runtime audio activity; use headed/manual proof when needed.
serialization in test tools; read proof/state as JSON strings before parsing in the test process.
For canvas-heavy games, include a pixel or screenshot check so a blank canvas cannot pass. For public Workflow/MCP behavior, supporting probes are not final: run the live chatbot ui-test when connector behavior changed.
If the game is not fully playable:
NOT_PLAYABLE_YET.browser autoplay, input mapping, restart loop, performance, or platform install capability.
STATUS.md if the blocker remains active.Typical sequence:
classic-game-design-test
-> frontend-ui-engineering
-> browser-testing-with-devtools
-> test-driven-development
-> code-review-and-qualityAdd security-and-hardening when handling user uploads, external archives, or remote game metadata. Add documentation-and-adrs when the solution changes the product architecture for browser-only game delivery.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.