Kybernesis PlatformBLOCK
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
The product monorepo for Kybernesis — the ecosystem layer that makes eve agent deployments enterprise-ready and uniquely Kybernesis. Eve is the plumbing; this is the moat.
Packages (all Apache-2.0, published to npm)
70ca2bd3fb63OBSERVED · 2026-10-12Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add voice --env ARCANA_API_KEY=${ARCANA_API_KEY} --env ARCANA_EVAL_API_KEY=${ARCANA_EVAL_API_KEY} --env ARP_AGENT_CREDENTIAL=${ARP_AGENT_CREDENTIAL} --env BLOB_READ_WRITE_TOKEN=${BLOB_READ_WRITE_TOKEN} -- npx -y @kybernesis/[email protected]{
"mcpServers": {
"voice": {
"command": "npx",
"args": [
"-y",
"@kybernesis/[email protected]"
],
"env": {
"ARCANA_API_KEY": "${ARCANA_API_KEY}",
"ARCANA_EVAL_API_KEY": "${ARCANA_EVAL_API_KEY}",
"ARP_AGENT_CREDENTIAL": "${ARP_AGENT_CREDENTIAL}",
"BLOB_READ_WRITE_TOKEN": "${BLOB_READ_WRITE_TOKEN}"
}
}
}
}Exposed tools (1)
1 read · 0 write · 0 destructive.
| Tool | Risk | Description |
|---|---|---|
Kyber | read | Kybernesis |
Trust audit
BLOCKgrade F · trust 44/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | WARN |
| L1 | Static analysis of the code | FAIL |
| L2 | Instruction surface (what it tells the agent) | FAIL |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (12 observation(s))
- Network
- declared (12 observation(s))
- Shell
- declared (5 observation(s))
- Dependencies
- not all pinned
- Secrets in source
- found
Findings (25)
if (/^\s*(?:(?:[A-Za-z_][A-Za-z0-9_]*=(?:"[^"]*"|'[^']*'|[^\s]+))\s+)*kyb-eval(?:\s|$)/.test(script)) {const direct = /^(\s*(?:(?:[A-Za-z_][A-Za-z0-9_]*=(?:"[^"]*"|'[^']*'|[^\s]+))\s+)*)eve\s+eval(?=\s|$)(.*)$/.exec(script);
const priv = join(keyDir, "id_ed25519");
re-read the credential per request: a Grok login expires in six hours and the
- **`eve start` does not read `.env.local`** the way `eve dev` does. Export it
kybernesis-buzz-0.9.8.tgz
kybernesis-manage-0.8.1.tgz
agentUrl: process.env.BUZZ_AGENT_URL ?? "http://127.0.0.1:8000",
BUZZ_AGENT_URL where the agent listens (default http://127.0.0.1:8000)
apiKey: "claude-subscription-local-proxy",
.vercelignore
import { queryRelay } from "../../src/query.js";export { default, environment } from "../../lib/reasoning-sandbox";const registryItem = JSON.parse(readFileSync(new URL("../../../registry/r/notify.json", import.meta.url), "utf8"));import { askingUser, notify, preview } from "../../src/index.js";node ../../scripts/check-eve-line.mjs package.json
process.stdout.write(" ✗ judge.autoevals.correctness\\n autoevals error: connect EHOSTDOWN 169.254.169.254:443\\nResults:\\n Failed: 1\\n");| `BUZZ_AGENT_URL` | Where the agent listens (default `http://127.0.0.1:8000`) |
agentUrl: "http://127.0.0.1:8000",
agentUrl: "http://127.0.0.1:8000",
@kybernesis/enterprise, jose
ai
@kybernesis/enterprise, nostr-tools, zod
zod, typescript, @types/node
@types/node, typescript
Gates applied: no_behavioural_pass.
70ca2bd3fb63full audit observations/trust-audit/mcp-server/kybernesisai__platform.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-12 | 70ca2bd3fb63 | BLOCK | F | 44 | first audit |
Questions
What tools does @kybernesis/arcana expose?
1 in total: 1 read-only, 0 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.
Is @kybernesis/arcana safe to connect to an agent?
No — not without reading the findings first. The audit graded it F (44/100) and found 5 critical or high issues in the source. Each one is listed on this page with the file and line it is on.
What credentials does @kybernesis/arcana need?
It reads ARCANA_API_KEY, ARCANA_EVAL_API_KEY, ARP_AGENT_CREDENTIAL, BLOB_READ_WRITE_TOKEN, BUZZ_AUTH_TAG, BUZZ_KEYFILE, COMPUTER_VNC_PASSWORD, DISCORD_BOT_TOKEN, EXE_API_TOKEN, EXE_SANDBOX_SSH_KEY, EXE_SANDBOX_SSH_KEY_PATH and GITHUB_TOKEN from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How current is this page?
The grade is for one exact copy of the source (70ca2bd3fb63), read on 2026-10-12. The repository is watched and re-audited when it changes.