All plugins

game-lab

Game Lab combines ModLab, Game Localization Studio, Sprites and experimental Playtest Pilot. ModLab works on a private game copy and invokes separately installed tools; localization drafts and exports RPG Maker MV/MZ text and the text of Unity Mono builds (JSON text assets, script-component strings and Assembly-CSharp string literals; patched copies of changed files only, the game folder is never written); localization companions estimate width and row budgets, merge and audit translation data, and proofread English; Sprites inspects and previews character sheets. Playtest Pilot uses approved X11 input, captures screen pixels within the selected window, stores private evidence frames and returns reduced previews and optional local OCR hints; for a failed check it can build a short captioned MP4 from those frames with your FFmpeg, or cut one from a recording file you point to (that asks for approval). On first use ModLab reads Steam's library list and app manifests to find installed games, lists the files in the chosen game folder, runs read-only hardware checks, and asks which modding route to use. It never fills a disk: a free-space reserve, a total cap, pruning of old revisions and a confirm-first cleanup tool. The Universal Modder and REA programs are the ones you set in the plugin settings or find on PATH (the model cannot name a program), and each toolkit call and each playtest check asks for your approval. Every program it starts gets a minimal environment, never your provider keys. The plugin makes no network calls of its own. Disclosure — Playtest Pilot sends synthetic key presses and clicks to the X11 window you select (via xdotool and Hermes computer_use, each check or batch approved), captures that window's pixels and stores evidence frames in the plugin's data folder; ModLab reads your Steam library and app-manifest files to list installed games, copies the game folder you choose into plugin data (with a disk cap), Game Localization Studio reads the RPG Maker data file or Unity game folder you point it at and copies the files that hold text into plugin data, writes a JSONL units file into the draft folder, and reads translations back from the JSONL file path named in the tool call (absolute path; symlinks, credential folders and the Hermes home are refused); and, after your approval of each call, runs the Universal Modder / REA programs set in the plugin settings or found on PATH.

flabanabba8Toolsv0.4.1Updated 20 hours ago

About

Game Lab combines ModLab, Game Localization Studio, Sprites and experimental Playtest Pilot. ModLab works on a private game copy and invokes separately installed tools; localization drafts and exports RPG Maker MV/MZ text and the text of Unity Mono builds (JSON text assets, script-component strings and Assembly-CSharp string literals; patched copies of changed files only, the game folder is never written); localization companions estimate width and row budgets, merge and audit translation data, and proofread English; Sprites inspects and previews character sheets. Playtest Pilot uses approved X11 input, captures screen pixels within the selected window, stores private evidence frames and returns reduced previews and optional local OCR hints; for a failed check it can build a short captioned MP4 from those frames with your FFmpeg, or cut one from a recording file you point to (that asks for approval). On first use ModLab reads Steam's library list and app manifests to find installed games, lists the files in the chosen game folder, runs read-only hardware checks, and asks which modding route to use. It never fills a disk: a free-space reserve, a total cap, pruning of old revisions and a confirm-first cleanup tool. The Universal Modder and REA programs are the ones you set in the plugin settings or find on PATH (the model cannot name a program), and each toolkit call and each playtest check asks for your approval. Every program it starts gets a minimal environment, never your provider keys. The plugin makes no network calls of its own. Disclosure — Playtest Pilot sends synthetic key presses and clicks to the X11 window you select (via xdotool and Hermes computer_use, each check or batch approved), captures that window's pixels and stores evidence frames in the plugin's data folder; ModLab reads your Steam library and app-manifest files to list installed games, copies the game folder you choose into plugin data (with a disk cap), Game Localization Studio reads the RPG Maker data file or Unity game folder you point it at and copies the files that hold text into plugin data, writes a JSONL units file into the draft folder, and reads translations back from the JSONL file path named in the tool call (absolute path; symlinks, credential folders and the Hermes home are refused); and, after your approval of each call, runs the Universal Modder / REA programs set in the plugin settings or found on PATH.

Game Lab

Game Lab

Work on games with Hermes: mod a game on a safe copy, translate its text without breaking it (RPG Maker MV/MZ and, new, Unity Mono builds), check character sprite sheets, and playtest it by pressing and holding keys and reporting honestly, with a short clip of each failed check.

New in 0.4.0: Game Localization Studio reads Unity games. Point localize_inspect at a Unity game folder (a Mono build with loose .assets files) and it drafts the text held in JSON text assets, in the string fields of script components and ScriptableObjects, and in the string literals of Assembly-CSharp.dll. localize_export writes patched copies of only the files that changed; your game folder is never written. Two new tools move a whole draft out to a JSONL file and back (localize_units_file, localize_import_translations), so a translator or a translation worker can do the wording. Checked on one real game's files by re-reading the output with two independent parsers; no game was started. See VERIFICATION.md.

New in 0.3.2: playtest clips. When a playtest_run check fails, you get a short captioned MP4 of the moment (3 seconds before to 1.5 after by default) saved beside the run's evidence, with its path and time range in the report. It is built from the frames the run already captured (shown as "frames from the run", a slideshow and not full-motion video), or cut from a screen recording you made yourself if you give its path and start time (new tool playtest_clips). Needs your own ffmpeg; without it the report is complete and says no clips were made. Nothing is recorded by the plugin and nothing is uploaded.

New in 0.2.2: playtest_run executes a checklist under one approval and returns a measured report. Playtest remains experimental: direct batch tests produced 0/4 complete correct SuperTux runs and 0/2 Packet Storm runs; no 0.2.2 Hermes sessions qualified or were attempted because the direct-test gate was not met. See VERIFICATION.md for limitations.

New in 0.2.1: Playtest Pilot adds an approved, single-call capture/input/measurement check and a magnified HUD crop. This part is experimental: 0/7 new SuperTux attempts and 0/5 Packet Storm attempts produced a complete correct report. See VERIFICATION.md for the causes and limits.

New in 0.2.0: a fourth part, Sprites, checks RPG Maker MV character sheets (and ModLab now tells you, before a mod is applied, if a character PNG in it does not fit the grid). After a mod is applied in a lab copy, or a translation is exported, Hermes now says plainly that the game can be playtested and offers it: you ask "Playtest it: ...". Nothing is started for you.

Why one plugin: mod it, translate it, check its sprites, then playtest it. Those are steps of the same job, so they live together.

Who it is for: someone who owns a PC game and wants Hermes to help change it, translate it and check that it still plays, without risking the original install.

The four parts

PartWhat it is forTools
ModLabCopy a game folder into a lab, apply mod ZIPs as revisions you can inspect and roll back, and drive Universal Modder and Reverse Engineer Anything for recon, assets, builds and tests. The original folder is only read.modlab_resources, modlab_toolkit, modlab_rea_native, modlab_create, modlab_inspect, modlab_apply, modlab_rollback, modlab_select_preserved
Game Localization StudioMake translation drafts for RPG Maker MV/MZ message text or for a Unity Mono build. Every string is addressed to its source; control codes, placeholders and (for RPG Maker) row widths are checked, and the result is exported as new files.localize_inspect, localize_get, localize_translate, localize_validate, localize_export, localize_list_drafts, localize_units_file, localize_import_translations
Playtest Pilot (experimental)Run an approved checklist and report measured verdicts, or explore a single input and record an observation.playtest_run, playtest_start, playtest_check, playtest_read_value, playtest_action, playtest_replay, playtest_observe, playtest_report, playtest_clips
SpritesInspect an RPG Maker MV character sheet PNG (single-character $ sheet or the ordinary eight-character sheet): grid, frame size, empty cells, alpha bounds. Preview one walk row as a separate animated GIF. Pairs with ModLab: before a mod that replaces character art is applied, check the sheet. Needs Pillow.sprite_inspect_mv, sprite_preview_mv

29 tools, one hook (pre_tool_call, which asks for approval before each batch, atomic check and legacy held key), eight skills (modlab, game-localization-studio, localization-overflow, game-translation-pipeline, localization-audit, translation-proofreading, playtest-pilot, sprite-contract) and one Desktop page for the translation drafts.

ModLab and Game Localization Studio are also published on their own. Playtest Pilot began from the stand-alone 0.3.0 plugin and now has bundle-specific atomic tools and approval handling. Sprites is the stand-alone Sprite Contract 0.1.0 plugin. The bundle also connects the parts through two text hints:

  • modlab_inspect adds a sprite_check note for PNGs under img/characters/ or img/sv_actors/ in the mod ZIP: does the sheet divide evenly into the MV grid for its filename. It only reports; it never blocks the mod.
  • modlab_apply and localize_export add a playtest line: the result is ready to be playtested, and how to ask ("Playtest it: ..."). The ModLab and localization skills tell Hermes to offer it. No game or playtest is ever started automatically.

At a glance

  • Install: hermes plugins install flabanabba8/game-lab, then hermes plugins enable game-lab.
  • Needs: Hermes 0.22 or later and a game you own or may translate.
    • Sprites and atomic playtest captures need Pillow (declared in the manifest, installed with --install-deps).
    • ModLab also needs the companion tools Universal Modder (um) and Reverse Engineer Anything (rea), installed by you, for the steps that use them. Its optional setting rea_native_tool names a native REA tool already registered in your Hermes.
    • Game Localization Studio needs RPG Maker MV/MZ JSON data files, or a Unity game folder built with Mono (it has <Game>_Data/Managed/) whose data is in loose .assets files. Hermes itself does the translating, or you fill in the JSONL file; no key or model is required by the plugin, and no extra Python package.
    • Playtest Pilot needs a model and provider that accept images in tool results, a Linux X11 (or XWayland) display, xdotool and Pillow with X11 capture support; optional tesseract reads potential panel text; optional ffmpeg (yours, on PATH) cuts the clips of failed checks. Hermes computer_use can list windows and remains necessary for legacy actions. The legacy scene hint optionally uses NumPy.
  • Platform: Linux is the only platform checked.
  • Internet: the plugin makes no network calls of its own: no telemetry, update checks or model calls. Any network use comes from the companion tools you choose to run with ModLab.
  • Keys and accounts: none read or stored.
  • Approvals: each atomic check and each legacy hold asks through this plugin’s approval box; other legacy actions use the Hermes computer_use approval policy.

Where things are stored

Each part keeps its data in its own Hermes plugin-data folder, so nothing is shared between them: game-lab-modlab (labs, revisions, toolkit logs), game-lab-localization (copies of the source file or, for Unity, of every file that holds text, which can be tens of megabytes per draft; drafts, units files and exports) and game-lab-playtest (playtest receipts, evidence frames and the <run id>-clips folders with the failure clips) and game-lab-sprites (preview GIFs). These are separate from the folders of the stand-alone plugins, so data written by those is not picked up; export or copy it over by hand if you need it. Data stays until you remove it.

Limits

  • The parts share no data. The only links are the two text hints above; you (or the agent) carry a game from one step to the next.
  • Sprites only knows the RPG Maker MV character grid. It cannot judge whether the artwork looks right or animates well, and its GIF is a frame-order preview, not the game. Only synthetic sheets were tested, no real artwork. The ModLab sprite_check looks at filenames under the two MV folders only.
  • ModLab works on a copy and copies whole revisions, so it needs disk space and time. It does not guarantee a copied game launches (launchers, DRM and saves outside the folder can get in the way), it does not install anything into your real game, and it blocks the publish and pull-request routes of the companion tools. No real mod and no game launch has been tested by us.
  • Game Localization Studio handles RPG Maker MV/MZ message text, choices and selected database fields, not scripts, images or pixel layout. A structural pass is not proof of in-game rendering, and the message row limit comes from you. Another engine needs its own importer.
  • Unity: only serialized-file format 22 (Unity 2020 and later), little-endian, loose files, up to 256 MiB each. Not covered: games packed into data.unity3d or asset bundles, IL2CPP code strings, text drawn into images, text the game builds at run time, and signed assemblies (refused). Component fields are found by their shape (a length-prefixed UTF-8 string that passes the text filter), not from type information, so a string field that is really an identifier can be offered for translation; a string equal to a game-object or asset name is blocked unless you allow it. No layout is checked: the validator only lists translations wider than their source. The game's fonts must contain the letters of the target language. A patched game has not been started by us.
  • Playtest Pilot measurements are visual heuristics, not proof of movement, a menu, a pause, or game state. The batch report can be wrong; inspect saved evidence before relying on it. A pause with animation and an unlabeled HUD value may be unverified. Capture timing can miss a short input; the result flags whether the during capture completed before release. Panel detection and OCR are fallible. Capture raises and focuses the window, then crops screen pixels; an always-on-top overlap can appear in that crop. Full-size check screenshots and records are saved as owner-only files in the playtest data folder; smaller JPEG previews are returned to the host, while HUD crops are returned magnified. See VERIFICATION.md for the live pass rate. Windows is unsupported.
  • Playtest clips are only as good as the run: a failed check gets the handful of frames saved for it (usually three, about a second apart), so the clip is a slow slideshow, not a recording of what the game did between them. Cutting from your own recording needs its start time to be right; a wrong clock gives a wrong moment. Runs made before 0.3.2 have no frame times and cannot be clipped. Clip tests used generated frames and videos and one real SuperTux run; see VERIFICATION.md.
  • Hermes Desktop gets one page, "Game Lab", with the editable translation draft view of Game Localization Studio (and its dashboard route; the web dashboard shows no "Game Lab" tab, because the route only serves the Desktop page). ModLab and Playtest Pilot have no page. The page was tested only in a headless browser fixture with a stub of the Hermes Desktop SDK, not in the real Desktop app.
  • The plugin bundles no game files, game data, media, SDKs or binaries.

Credits and licences

  • Sprites follows RPG Maker MV's asset standards for the character grid; it bundles no RPG Maker files.
  • Universal Modder and Reverse Engineer Anything are separate projects used by ModLab through their command-line tools. Their authors and licences apply; nothing from them is bundled.
  • RPG Maker MV is a product of its owners. The official MV event interpreter and the TextManager source from the public rpgtkoolmv/corescript repository were used in test fixtures only; they are not bundled, and Game Localization Studio is not affiliated with or endorsed by RPG Maker.
  • Playtest Pilot uses xdotool, Pillow and optionally Tesseract; legacy tools also use Hermes computer_use, ffmpeg and optionally NumPy; clips use ffmpeg and Pillow. These executables are not bundled.
  • Licence: GPL-3.0-or-later (see LICENSE).

AI assistance

Written with AI coding assistants (Anthropic Claude and OpenAI Codex agents) under the maintainer's direction. What was actually tested is listed in VERIFICATION.md.

Reading further

Bundled wiki, tool contracts, verification record.

README from the reviewed commit 34c35d9; it updates when the author re-pins.

More by flabanabba8

6 plugins by flabanabba8 in the catalog.

flabanabba8live2d-riggerCreate a provisional Live2D character from a brief with complete image parts, local cutout, placement, an experimental Umamo/JDK rig and Core motion review; also inspect and adjust a copy of an existing model. XY head turns, independent limbs and polished eyes are not provided. Disclosure — runs the Java (JDK) and Kotlin/Umamo classpath, the Python with Playwright and the Cubism SDK viewer folder that you set in the plugin settings (the model cannot name a program, classpath or command), plus FFmpeg and a local headless browser against a viewer served only on a temporary loopback port, each with a minimal environment; on first use it scans PATH and the hardware (read-only) and asks which routes to use; opening an existing model copies its folder into the plugin's data directory; on request it downloads official Live2D sample archives from cubism.live2d.com after you agree to Live2D's licence; the host image tool may use its configured provider; no Cubism SDK, Core, viewer, sample model, layers or textures are bundled, and the catalog thumbnail contains character artwork.Communitytools1flabanabba8case-trailTrace case-history and treatment research to actual opinion passages; analyze supplied sources without a required repository. Disclosure — Sends case search queries and Justia page URLs through the web search/extract provider you configured in Hermes, and can fetch one public OpenJurist case page you select after a robots.txt check (OpenJurist sees the URL and your IP). Reads local opinion files at paths given in chat, refusing kernel and device files, the Hermes home and key stores; the plugin setting opinion_folders can limit it to folders you list. Stores fetched text, review receipts and rendered pages privately in its plugin data folder. Novelty research tool, not legal advice; unaffiliated with Justia or OpenJurist.Communitywebflabanabba8creator-studioMake media with Hermes from one request: researched lesson videos, two-host podcast episodes, captions for any recording, narrated step-by-step tutorials, and clips cut from a YouTube video or stream by reading its transcript. Bundles five plugins (YouTube Studio, Podcast Studio, Caption Studio, Tutorial Creator, Clip Studio) into one entry. Disclosure — Research pages are fetched through the Camoufox browser you set in the plugin settings, in a fresh anonymous profile (public addresses only). On first use it lists the voices and speech recognizers this computer has and asks you which to use; it never picks one for you. If you choose Gemini narration, each spoken line is sent to Google, either to Vertex AI (aiplatform.googleapis.com) with a Google sign-in that Hermes holds for your own OAuth client (hermes auth add creator-studio:google; the plugin never sees the token), or to generativelanguage.googleapis.com with the Gemini API key in this plugin's settings (GEMINI_API_KEY, the same variable Hermes' own Gemini provider uses), behind a spend guard (default 0.50 USD). To find speech servers it sends read-only GET requests (/v1/models, /openapi.json) to 127.0.0.1 only, on ports 2022, 5000, 5002, 7860, 8000, 8001, 8020, 8080, 8880 and 9000, plus any loopback address you set. Caption Studio runs the recognizer command you set in the plugin settings, or uploads audio to the HTTPS speech service you set (public addresses only) or to Groq (api.groq.com) with your key, only when you allow uploads; the model cannot set commands, endpoints, keys or upload consent. Programs it starts get a minimal environment (paths, locale, display/audio session and proxy settings), never your provider keys. Clip Studio runs the yt-dlp you installed (or set in the plugin settings) and makes requests to YouTube only when you give it a link: the video page and one caption track, then the audio track if you ask for speech recognition, the sections you choose to cut, and, if you ask what chat reacted to, the public chat replay of a finished stream (other people's messages and display names; kept in private plugin storage, the raw download deleted once counted, names kept only for paid messages). It never uses browser cookies or a YouTube sign-in. Recognizing a whole recording continues in a background thread inside Hermes until its pieces are done. Clips, joined videos, titles, descriptions and thumbnails are written to plugin storage. Runs FFmpeg and ffprobe, an installed speech program if you choose one (espeak-ng, flite, piper), and for screen tutorials xwininfo, xprop, xdotool, pactl and xclip; the recorder starts one detached FFmpeg worker per take, and tutorial steps click and type through Hermes computer_use, so its approvals apply. Stores projects privately in plugin storage. Nothing is published.Communitytoolsflabanabba8pagecast-directorEditable cast narration with locked voices, source passage IDs, resumable synthesis and selective regeneration. Disclosure — Stores books and generated audio privately in its plugin data. On first use it lists the speech options on this computer and asks which to use; it never picks one for you, and online engines are blocked if you say no online services. To find local speech servers it sends GET /v1/models only to this computer's loopback address (ports 8000, 8880, 5002, 8020 and 8080, plus any local address you set; no redirects, 1 second timeout). Every program it starts gets a minimal environment (PATH, HOME, LANG), never your provider keys. Voices are local (an installed speech program such as espeak-ng, Flite, Piper or say, or a speech server already running on this computer) unless you choose hosted Inworld, ElevenLabs or Google Gemini, which send passage text and voice choices to that provider with your own key and use your credits; the optional Gemini Vertex route uses a Google sign-in that Hermes holds for your own OAuth client (hermes auth add pagecast-director:google; the plugin never sees the token), sent only to aiplatform.googleapis.com. Can render a short captioned video clip (audiogram) of a finished chapter span into plugin data, with a cover image you point to. Runs local FFmpeg/ffprobe and adb; mpv only plays a preview aloud when you approve it. The optional Android tools check the author's GitHub releases (flabanabba8/PageCast) for the PageCast app and can download the latest APK (checked against that release's own SHA256SUMS); installing it on a phone with adb, replacing a locked voice and playing a preview aloud each ask for your approval (one pre_tool_call hook). No background processes beyond a render or preview you start.Communityvoice

Developer Mode

Explore built-in Hermes toolsets and actions.

HermesAgent

Developer?

Create a Hermes extension for Desktop or agent workflows.

Developer Section