alibabacloud-cms-manage

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly fetches and renders/executes remote onboarding code at runtime — e.g. it calls aliyun cms2 integration addon get to retrieve addon code templates that are rendered into instructions, it instructs downloading agent binaries from URLs like "http://arms-apm-{regionId}.oss-{regionId}.aliyuncs.com/.../AliyunJavaAgent.zip" (wget + unzip), and it even documents installer patterns that run remote scripts such as curl -fsSL ... | bash, all of which are runtime external dependencies that can execute remote code or directly control agent instructions.

HIGH W008: Secret detected in skill content (API keys, tokens, passwords).

• Secret detected (high risk: 1.00). I scanned the skill text for literal high-entropy credentials. I found a concrete token in the example apm configuration response:
• In references/apm.md (Step 3 — "Example real response") the JSON contains: "entryPointInfo": { "authToken": "awy7aw18hz@26*44b70", ... }

This value is a non-placeholder, random-looking token that the document even maps to LicenseKey / authToken and calls it sensitive ("Agent authentication token (sensitive)"). Therefore it meets the definition of a secret (high-entropy credential that grants reporting access).

Ignored items / why not flagged:

• service identifiers like "pid": "awy7aw18hz@9269550ea2c2be0" and "serviceId": "awy7aw18hz@645ab0bc177a46e87c7f1" are resource IDs/identifiers, not authentication secrets — not flagged.
• workspace names, requestId, endpoints, and account/workspace numeric IDs are not secrets.
• All placeholder patterns ({LicenseKey}, {workspace}, YOUR_API_KEY, etc.) and env var names are intentionally ignored per the rules.

Conclusion: there is an actual credential-like token present in the example response (entryPointInfo.authToken), so this counts as a real secret exposure in the document.

MEDIUM W021: Hidden or invisible Unicode characters detected (potential obfuscation or prompt injection).

• Hidden Unicode characters detected (1 type(s) found)