OGAD
Off Grid AI — private, on-device AI. Run open models (text, vision, image, voice) locally through one OpenAI-compatible gateway. No cloud, no accounts, no API keys.
Warning. Serious findings were identified. Review the full report before connecting. Grades reflect the full trust pyramid: code, verification depth, and reputation. New projects cap at C until adoption is earned.
How we verified
Verified is a snapshot. Live keeps it current, and builds your track record.
⚡ Connect GitHub → continuous verification on every pushwhy connect →Who stands behind it
off-grid-ai
Source: github_repo_search
Claim it to get a verified publisher badge, a free copy of our full audit findings, and direct contact for any high-priority issues we find. Or connect your repo for our deepest verification, Live Monitored: read-only, revoke anytime. What we access →
Install from
The grade above is for the source repository. Registries can serve a different version, so we mark the ones we were not able to read.
These names and descriptions are the publisher's own, read from the source code. We print them as written. Our assessment is the findings above, not this list.
generate_textGenerate text with the local LLM (Off Grid AI). Supports an optional system prompt.
describe_imageVision: analyze an image (data URL, http(s) URL, or local path) with the local VLM.
generate_imageText-to-image with the local diffusion model. Returns a PNG.
edit_imageRepaint an input image guided by a prompt (img2img) with the local diffusion model.
transcribe_audioTranscribe an audio file (data URL, http(s) URL, or local path) with the local whisper model.
text_to_speechSynthesize speech from text with the local Kokoro voice model. Returns WAV audio.
embedEmbed text with the local all-MiniLM-L6-v2 model (384-dim). Returns the vector as JSON.
calendar_create_eventcalendar_list_eventsreminders_createreminders_listcontacts_searchmessages_sendmail_sendopen_urlcomputer_useDisclosed vulnerabilities in this server's declared npm dependencies (via OSV). Whether each is reachable depends on the installed versions.
sharp: Vulnerabilities in libheif: GHSA-g89c-p67h-r497 and GHSA-2jg2-4ch7-h545
ws: Memory exhaustion DoS from tiny fragments and data chunks
Electron: Use-after-free in offscreen child window paint callback
Electron: Use-after-free in WebContents fullscreen, pointer-lock, and keyboard-lock permission callbacks
Electron: Sandboxed iframe can bypass the allow-popups restriction via the OpenURL navigation path
ELECTRON_RENDERER_URLOFFGRID_DATA_DIROFFGRID_DIAGNOSTIC_LOGOFFGRID_FORCE_COREOFFGRID_SEEDOFFGRID_SERVER_ONLYOFFGRID_USER_DATASystemRootDependencies
59 runtime dependencies (38 dev), 1 flagged: @playwright/mcp
Tool annotations
No tools have read-only/destructive annotations
Add readOnlyHint or destructiveHint annotations to every tool so hosts can warn users before invoking.
All four hints declared on every tool
16/16 tools missing one or more hints — generate_text (missing: readOnlyHint, destructiveHint, idempotentHint, openWorldHint); describe_image (missing: readOnlyHint, destructiveHint, idempotentHint, openWorldHint); generate_image (missing: readOnlyHint, destructiveHint, idempotentHint, openWorldHint), +13 more. OpenAI's directory rejects tools where any of the four hints are missing or non-boolean.
For every tool, set all four hints (readOnlyHint, destructiveHint, idempotentHint, openWorldHint) to explicit true/false values that match the handler’s actual behaviour.
Destructive tools are labelled
2 tools perform destructive updates without destructiveHint — edit_image deletes at line 228 (fs.promises.unlink(initImage)); transcribe_audio deletes at line 251 (fs.promises.unlink(p))
Add destructiveHint:true to any tool whose handler calls .delete(), .upsert(), .update(), unlink, rm, DELETE, DROP, REPLACE INTO, or any operation that overwrites existing data.
Tool test coverage
11/16 tools referenced in tests (69%)
Write tests that reference each tool by name so every tool has at least one test.
Shell command execution
13 calls in production code run through a shell (src/main/sd-server.ts:286, src/main/sd-server.ts:287, src/main/transcription/whisper-server.ts:185)
Prefer library functions over shell-outs. If you must shell out, ensure all inputs are properly escaped.
Readable source code
1 file are minified or bundled, which is usually build output rather than concealment
Ship unminified, readable source.
No arbitrary install scripts
Has postinstall/preinstall script — runs arbitrary code on npm install
Remove postinstall/preinstall hooks unless they’re essential.
Production dependencies are patched
0 critical, 2 high severity in production deps — sharp@0.35.3 (high), ws@8.18.3 (high)
Run npm audit fix, or upgrade the affected packages to a non-vulnerable version.
Dev dependencies
15 critical/high in dev-only deps (does not ship to users)
Upgrade dev dependencies when convenient.
Dependency freshness
11/59 production deps abandoned (no release in 2+ years): kdbxweb@2022-06-19 (4.2y), node-machine-id@2023-05-09 (3.4y), tweetnacl@2023-04-12 (3.4y)
Claim the listing to review these findings one by one and send us a correction where you disagree, straight to the team. Claiming also means we tell you when the grade moves, and reach you first if we find anything urgent.
[](https://m8ven.ai/mcp/off-grid-ai/ogad)?variant=verified from the URL.Vetting this one by hand? Tool Check is an MCP that scores other MCPs. Add it once and ask Claude, ChatGPT, or any MCP client to grade a server, surface CVEs, check the publisher, and suggest safer alternatives — before you install.
https://m8ven.ai/api/mcp/tool-check