registry / @one-source/mcp / 5.23.1
@one-source/mcp@5.23.1
Blockchain data, Deepstate and Standard Reserve markets, live chain queries, and API docs
AI Security Review
scanned 2d ago · by lpm-firewall-aiThe flagged API key is submitted through the OAuth connect flow and validated against OneSource’s API before an auth code is issued. Runtime analytics is enabled by default, but it reports usage metadata, logs its endpoint at startup, and can be disabled with ONESOURCE_ANALYTICS=false; no credential theft or unrelated collection was found.
Static reason
High-risk behavior combination matched malicious policy.
Trigger
The key is sent only when a user submits the OAuth connect form. Analytics events are generated when tools run and flushed periodically or after enough events accumulate.
Impact
No malicious impact identified from the reviewed source. Usage metadata leaves the runtime when analytics is enabled.
Mechanism
The key is checked against the configured OneSource API. Analytics sends tool usage fields to the OneSource analytics service; ONESOURCE_ANALYTICS=false disables collection.
AI rationale
The flagged API key is submitted through the OAuth connect flow and validated against OneSource’s API before an auth code is issued. Runtime analytics is enabled by default, but it reports usage metadata, logs its endpoint at startup, and can be disabled with ONESOURCE_ANALYTICS=false; no credential theft or unrelated collection was found.
Network endpoints3
api.onesource.io/api/chain/chain-id (or ONESOURCE_BASE_URL override)1s-analytics.vercel.app/api/eventsregistry.npmjs.org/@one-source/mcp/latest
Decision evidence
public snapshotAI called this Clean at 91.0% confidence as Benign with low false-positive risk.
Evidence for AI clean decision
- The prepare script runs TypeScript compilation; the package entry points are the server factory and CLI.
- The connect-submit route accepts an API key as user input and validates it against OneSource’s API before issuing an auth code.
- The key validation request sends the candidate to the configured OneSource API base URL, defaulting to api.onesource.io.
- Runtime analytics reports tool names, input parameter names, timing, status, and response size rather than tool input values or responses.
- Analytics is sent to the OneSource analytics endpoint by default; startup logs its status and endpoint, and ONESOURCE_ANALYTICS=false disables it.
- The payment wallet identifier included in analytics is a truncated SHA-256 hash.
Evidence against
- No automatic credential discovery or transmission to an unrelated host was identified.
- The flagged API key flow requires a connect-submit request containing the user's key and returns an authorization code after validation.
- Analytics events contain usage metadata, not tool arguments or output; analytics can be disabled with an environment variable.
Behavioral surface
Source & flagged code
4 flaggeddist/connect-page.jsView file
9const MIN_KEY_FAIL_MS = 250;
L10: // Accepts the base64url output of randomBytes(16) used as internalState in oauth.ts.
L11: const STATE_RE = /^[A-Za-z0-9_-]{1,100}$/;
...
L27: res.writeHead(status, JSON_HEADERS);
L28: res.end(JSON.stringify({ error }));
L29: }
...
L31: try {
L32: const base = (process.env.ONESOURCE_BASE_URL ?? 'https://api.onesource.io').replace(/\/+$/, '');
L33: const r = await fetch(`${base}/api/chain/chain-id?network=ethereum`, {
Critical
Credential Exfiltration
Source appears to send environment or credential material to an external endpoint.
dist/connect-page.jsView on unpkg · L99Trigger-reachable chain: manifest.bin -> dist/cli.js -> dist/connect-page.js
L9: const MIN_KEY_FAIL_MS = 250;
L10: // Accepts the base64url output of randomBytes(16) used as internalState in oauth.ts.
L11: const STATE_RE = /^[A-Za-z0-9_-]{1,100}$/;
...
L27: res.writeHead(status, JSON_HEADERS);
L28: res.end(JSON.stringify({ error }));
L29: }
...
L31: try {
L32: const base = (process.env.ONESOURCE_BASE_URL ?? 'https://api.onesource.io').replace(/\/+$/, '');
L33: const r = await fetch(`${base}/api/chain/chain-id?network=ethereum`, {
Critical
Trigger Reachable Dangerous Capability
A package entrypoint or install-time lifecycle script reaches a source file with blocking dangerous behavior.
dist/connect-page.jsView on unpkg · L9Findings
2 Critical2 High3 Medium6 Low
CriticalCredential Exfiltrationdist/connect-page.js
CriticalTrigger Reachable Dangerous Capabilitydist/connect-page.js
HighTrigger Reachable Credential Exfiltrationdist/connect-page.js
HighSemantic Analysis Limiteddist/register-docs-tools.js#virtual:normalized:round1
MediumNetwork
MediumEnvironment Vars
MediumStructural Risk Force Deep Review
LowNon Install Lifecycle Scripts
LowScripts Present
LowEval
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings
Affected versions and remediation
This report applies to @one-source/mcp@5.23.1.
See all versions of @one-source/mcp for other recorded verdicts.
- Review the evidence and your use of @one-source/mcp@5.23.1 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.