Browser Session Manager

AdvisoryAudited by Static analysis on Apr 30, 2026.

Overview

No suspicious patterns detected.

Findings (0)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

If used with the wrong URL or session file, authentication tokens could be replayed or exposed in a way that gives access to the user's Jimeng account session.

Why it was flagged

The helper reads an arbitrary session JSON file and applies all included cookies and storage to a browser session for a caller-supplied URL, without a Jimeng domain allowlist or filtering of sensitive tokens.

Skill content
async function applySessionData(url, sessionJsonPath, options = {}) ... const sessionData = JSON.parse(fs.readFileSync(sessionJsonPath, 'utf8')) ... await context.addCookies(cookies) ... localStorage.setItem(key, value)
Recommendation

Restrict the helper to explicit Jimeng domains, validate cookie domains before use, filter only required keys, and require clear user approval before loading exported session data.

What this means

Following this guidance may violate service rules, trigger account restrictions, or enable automation beyond normal user-controlled usage.

Why it was flagged

The guidance explicitly recommends alternate IP/proxy use when the service rate-limits automated access, which can bypass provider controls and amplify automated activity.

Skill content
**Rate limited:**
- Add delays between requests
- Use different IP/proxy if heavily rate-limited
Recommendation

Remove rate-limit bypass guidance and instead tell users to respect provider limits, reduce request volume, and stop automation when rate-limited.

What this means

Running the script may submit generation requests and spend account credits without an additional in-script confirmation step.

Why it was flagged

The script can automatically click the logged-in Jimeng generation submit button, which is purpose-aligned but performs an account action that may consume credits.

Skill content
const submitBtn = await page.locator('button[class*="submit"], button[class*="send"], [class*="circle"]:has(svg), button:has([class*="arrow"])').last();
await submitBtn.click({ timeout: 10000 });
Recommendation

Add an explicit confirmation or dry-run mode before submitting jobs, especially when account credits or paid usage may be affected.

What this means

If the visited website or browser is compromised, disabling the sandbox can increase local system risk.

Why it was flagged

The skill launches Chromium against external web content with sandbox protections disabled, which is common in some containers but weakens browser isolation.

Skill content
const browser = await chromium.launch({
    headless,
    args: ['--no-sandbox', '--disable-setuid-sandbox']
  });
Recommendation

Avoid disabling the browser sandbox unless the runtime requires it, and document the risk when these flags are necessary.

What this means

Users must supply/install dependencies themselves and cannot verify dependency versions from the provided skill metadata.

Why it was flagged

The package provenance is limited and there is no install spec, even though the artifacts include runnable Node/Playwright scripts and documentation requiring external tools.

Skill content
Source: unknown
Homepage: none
No install spec — this is an instruction-only skill.
Recommendation

Provide a clear install spec or package manifest with pinned dependencies and a project homepage or source repository.