registry / @validate.qa/runner / 1.0.102
@validate.qa/runner@1.0.102
validate local test runner - execute Playwright tests on your machine, connected to validate.qa cloud.
AI Security Review
scanned 3d ago · by lpm-firewall-aiThe flagged browser code implements scoped Google sign-in for local QA, while the generated Playwright code and write guard fit the stated runner purpose. I found no automatic credential theft, unrelated account action, or install-time attack; access to the saved Google session by tests is disclosed.
Static reason
High-risk behavior combination matched malicious policy.; source matched previously finalized malicious package; routed for review; previous stored version diff introduced dangerous source
Trigger
Google sign-in requires an explicit login setup and is used when a test app starts its sign-in flow.
Impact
No malicious impact was identified. The saved Google session is available to tests for allowed sites, a risk the package discloses.
Mechanism
The runner brokers sign-in for named sites and applies a browser write guard; generated Playwright code is checked before use.
AI rationale
The flagged browser code implements scoped Google sign-in for local QA, while the generated Playwright code and write guard fit the stated runner purpose. I found no automatic credential theft, unrelated account action, or install-time attack; access to the saved Google session by tests is disclosed.
Network endpoints1
accounts.google.com
Decision evidence
public snapshotAI called this Clean at 91.0% confidence as Benign with low false-positive risk.
Evidence for AI clean decision
- The manifest describes a local Playwright test runner connected to the validate.qa cloud.
- The only lifecycle script is prepublishOnly, which runs the build.
- Google sign-in is exposed as an explicit google-login command; the flow requires named sites and refuses shared cloud runners.
- The Google broker flow is triggered when the app starts sign-in and uses a run-provided relay URL and token.
- The dynamic Function use checks generated Playwright code for syntax; the surrounding code rejects code over 200,000 characters.
- The browser write guard can reject blocked data-changing fetch requests.
Evidence against
- The package warns that tests from projects whose runs reach this machine can use the saved Google sign-in for allowed sites, and that programs running as the user may access it.
Behavioral surface
ChildProcessCryptoDynamicRequireEnvironmentVarsFilesystemNetworkShell
Source & flagged code
17 flaggeddist/discovery-mcp-server.jsView file
15};
L16: var __commonJS = (cb, mod) => function __require() {
L17: try {
Medium
Dynamic Require
Package source references dynamic require/import behavior.
dist/discovery-mcp-server.jsView on unpkg · L15dist/run-write-guard.jsView file
138ACTION_BATCH_MAX_SIZE = 10;
L139: MOBILE_SCREENSHOT_INTERVAL_MS = typeof process !== "undefined" ? parseInt(process.env?.MOBILE_SCREENSHOT_INTERVAL_MS ?? "500", 10) : 500;
L140: MOBILE_ACTION_BATCH_MAX_SIZE = 5;
...
L314: return false;
L315: if (typeof f.base64 !== "string" || f.base64.length === 0)
L316: return false;
...
L478: "javascript:",
L479: "data:",
L480: "vbscript:",
...
L1162: ``,
L1163: `main().catch((err) => { console.error(err); process.exitCode = 1; });`
L1164: ].join("\n");
Critical
Browser Credential Phishing
Source appears to collect browser login credentials for exfiltration.
dist/run-write-guard.jsView on unpkg · L138Findings
3 Critical14 High7 Medium7 Low
CriticalBrowser Credential Phishingdist/run-write-guard.js
CriticalRemote Response Code Executiondist/chunk-CBVHPYVJ.mjs
CriticalPrevious Version Dangerous Deltadist/chunk-DN3O6EYT.mjs
HighChild Process
HighShell
HighEval
HighBrowser Session Account Hijackdist/run-write-guard.js
HighCross File Remote Execution Contextdist/chunk-DN3O6EYT.mjs
HighSemantic Analysis Limiteddist/discovery-mcp-server.mjs#virtual:normalized:round1
HighKnown Malware Source Similaritydist/credential-mcp-server.js
HighKnown Malware Source Similaritydist/tar-extract-child.js
HighKnown Malware Source Similarity
Affected versions and remediation
This report applies to @validate.qa/runner@1.0.102.
See all versions of @validate.qa/runner for other recorded verdicts.
- Review the evidence and your use of @validate.qa/runner@1.0.102 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.