registry / Credential exfiltration / @tapi-dev/sdk / 0.3.12

@tapi-dev/sdk@0.3.12

## Live Railway server diagnostics (operator MCP)

AI Security Review

scanned 21d ago · by lpm-firewall-ai

The active CLI forwards existing service credentials to an unrestricted configured endpoint, warranting a warning under the credential-forwarding policy. Inspected evidence does not establish malicious intent or justify blocking publication.

Static reason

High-risk behavior combination matched malicious policy.; source matched previously finalized malicious package; routed for review

Trigger

Run a credential-backed command such as sessions with --server or a workspace apiBaseUrl override.

Impact

A configured recipient outside the intended service could receive the user's service token.

Mechanism

The CLI obtains cached service credentials, combines their ID token with the configured API URL, and sends it as a bearer token.

AI rationale

Evidence

dist/auth.jsdist/cli.jsdist/workflow-cli.jsdist/dev-context.jsdist/index.jsdist/client.jsauth.jsontapi.config.json

Network endpoints1

determined-motivation-production.up.railway.app

Decision evidence

public snapshot

AI called this Suspicious at 96.0% confidence as Dangerous Capability with medium false-positive risk.

Evidence for warning

package.json declares no registry-install lifecycle hooks.
dist/auth.js reads an existing cached ID token.
dist/cli.js activates credential-backed workflow commands on user invocation.
dist/workflow-cli.js accepts a server override or workspace API URL and pairs it with cached credentials; dist/dev-context.js imposes no destination restriction.
dist/index.js connects the CLI client to dist/client.js, which sends the token in an Authorization header.

Evidence against

The default API destination is package-aligned.
The observed forwarding requires an explicit CLI command; no hidden recipient or malware intent was established.

Behavioral surface

Source

ChildProcessCryptoDynamicRequireEnvironmentVarsEvalFilesystemNetworkWebSocket

Supply chain

HighEntropyStringsUrlStrings

Manifest

Source & flagged code

12 flagged

dist/auth.jsView file

Lines 1-26javascript

1import { randomUUID } from "node:crypto";

2import { existsSync, readFileSync } from "node:fs";

3import { mkdir, rename, rm, writeFile } from "node:fs/promises";

4import { homedir } from "node:os";

5import { dirname, join } from "node:path";

6const DEFAULT_FIREBASE_API_KEY = "AIzaSyCDZR8lWyVQcWYfFdNZa4vuL4IWEC0h6gE";

High

High Secret

Package contains a high-severity secret pattern.

dist/auth.jsView on unpkg · L6

High

Secret Pattern

Google API key in dist/auth.js

dist/auth.jsView on unpkg · L6

7const DEFAULT_EXPIRY_SKEW_MS = 60_000;

8export function getTapiDataDir() {

9 if (process.platform === "win32") {

10 return join(process.env.LOCALAPPDATA || join(process.env.USERPROFILE || homedir(), "AppData", "Local"), "Tapi");

11 }

12 return join(process.env.XDG_DATA_HOME || join(homedir(), ".local", "share"), "tapi");

13}

14export function getTapiAuthCachePath(options = {}) {

15 return options.cachePath || join(getTapiDataDir(), "auth.json");

16}

17export function readTapiCredentials(options = {}) {

18 const path = getTapiAuthCachePath(options);

19 if (!existsSync(path))

20 return null;

21 try {

22 const value = JSON.parse(readFileSync(path, "utf8"));

23 const credentials = {

24 uid: String(value.uid || "").trim(),

25 idToken: String(value.id_token || "").trim(),

26 refreshToken: String(value.refresh_token || "").trim(),

dist/workflow-cli.jsView file

Lines 1-23javascript

1import { mkdir, writeFile } from "node:fs/promises";

2import { loadWorkspace } from "./workspace.js";

3import { resolve, join } from "node:path";

4import { spawn } from "node:child_process";

dist/cli.jsView file

Lines 7-47javascript

7import { mkdir, readFile, readdir, rename, rm, stat, statfs, unlink, writeFile } from "node:fs/promises";

8import { homedir } from "node:os";

9import { basename, dirname, join, resolve } from "node:path";

10import { performance } from "node:perf_hooks";

dist/studio-dev.jsView file

Lines 1-21javascript

1import { spawn } from "node:child_process";

2import { createServer } from "node:net";

3import { existsSync, readFileSync } from "node:fs";

Critical

Reverse Shell

Source matches reverse-shell style process and socket wiring.

dist/studio-dev.jsView on unpkg · L1

Critical

dist/workflow-source.js#virtual:normalized:round1View file

•stage = ast_semantic_analysis; reason = ast_parse_error; limitedFiles = 1

High

Semantic Analysis Limited

A bounded semantic-analysis stage reached its safety limit; remaining detectors completed, but this package requires AI review.

dist/workflow-source.js#virtual:normalized:round1View on unpkg

Source & flagged code

12 flagged · loading source

dist/auth.jsView file

6patternName = google_api_key severity = high line = 6

High

High Secret

Package contains a high-severity secret pattern.

dist/auth.jsView on unpkg · L6

6patternName = google_api_key severity = high line = 6

High

Secret Pattern

Google API key in dist/auth.js

dist/auth.jsView on unpkg · L6

dist/workflow-cli.jsView file

3import { resolve, join } from "node:path"; L4: import { spawn } from "node:child_process"; L5: import { fileURLToPath } from "node:url";

High

Child Process

Package source references child process execution.

dist/workflow-cli.jsView on unpkg · L3

Findings

2 Critical10 High4 Medium7 Low

CriticalReverse Shelldist/studio-dev.js

CriticalTrigger Reachable Dangerous Capabilitydist/studio-dev.js

HighHigh Secretdist/auth.js

HighChild Processdist/workflow-cli.js

HighCredential Redirect Persistencedist/cli.js

HighSandbox Evasion Gated Capabilitydist/studio-dev.js

HighCross File Remote Execution Contextdist/workflow-cli.js

HighSemantic Analysis Limiteddist/workflow-source.js#virtual:normalized:round1

HighKnown Malware Source Similaritydist/cli.js

HighKnown Malware Source Similaritydist/studio-dev.js

HighSecret Patterndist/auth.js

High

Affected versions and remediation

This report applies to @tapi-dev/sdk@0.3.12.

See all versions of @tapi-dev/sdk for other recorded verdicts.

Review the evidence and your use of @tapi-dev/sdk@0.3.12 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: 2026-10-01 18:43:17.380Z (UTC).

Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.