Skills relational-database-mcp-cloudbase Audit History

📦

Audit History

relational-database-mcp-cloudbase - 2 audits

Version comparison

Capability and finding changes across audited versions, newest first.

Version	Date	Result	Review items	Change vs previous
v2 Latest	Jul 9, 2026, 04:44 PM	1 confirmed	4	No capability change
v1	Jul 9, 2026, 04:44 PM	1 confirmed	4	Baseline

Audit version 2 Latest

Jul 9, 2026, 04:44 PM

Most static command, filesystem, and blocker findings are false positives caused by Markdown formatting and static documentation references. The hardcoded remote fallback URLs are a confirmed low-risk network concern, and the remote instruction fallback is a medium supply-chain trust concern.

Files scanned

211

Lines analyzed

Review items

False positives ignored

Confirmed security concerns (1)

Medium

SKILL.md:12-15 SKILL.md:27-30

Remote Instruction Fallback Expands Trust Boundary

The skill tells agents to use published raw URLs as fallback sources for main and sibling skill instructions. Those instructions are outside the packaged audited artifact and could change after review.

The remote fallback URLs and sibling skill instructions are explicit in SKILL.md. I found no malicious text, but the trust-boundary expansion is real.

Capability review items (4)

These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.

Low

SKILL.md:12

Hardcoded URL

- CloudBase main entry: `https://cnb.cool/tencent/cloud/cloudbase/cloudbase-skills/-/git/raw/main/sk

This is a hardcoded remote documentation URL used as a fallback source. It does not exfiltrate data, but it can make agents consume instructions outside the audited package.

Low

SKILL.md:13

Hardcoded URL

- Current skill raw source: `https://cnb.cool/tencent/cloud/cloudbase/cloudbase-skills/-/git/raw/mai

This is a hardcoded remote documentation URL used as a fallback source. It does not exfiltrate data, but it can make agents consume instructions outside the audited package.

Low

SKILL.md:29

Hardcoded URL

- Web application integration -> `../relational-database-web/SKILL.md` (standalone fallback: `https:

This is a hardcoded remote documentation URL used as a fallback source. It does not exfiltrate data, but it can make agents consume instructions outside the audited package.

Low

SKILL.md:30

Hardcoded URL

- Raw HTTP database access -> `../http-api/SKILL.md` (standalone fallback: `https://cnb.cool/tencent

This is a hardcoded remote documentation URL used as a fallback source. It does not exfiltrate data, but it can make agents consume instructions outside the audited package.

Risk Factors

⚙️ External commands (50)

🌐 Network access (4)

SKILL.md:12 SKILL.md:13 SKILL.md:29 SKILL.md:30

📁 Filesystem access (2)

SKILL.md:29 SKILL.md:30

Audited by: codex

Audit version 1

Jul 9, 2026, 04:44 PM

Files scanned

211

Lines analyzed

Review items

False positives ignored

Confirmed security concerns (1)

Medium

SKILL.md:12-15 SKILL.md:27-30

Remote Instruction Fallback Expands Trust Boundary

The remote fallback URLs and sibling skill instructions are explicit in SKILL.md. I found no malicious text, but the trust-boundary expansion is real.

Capability review items (4)

These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.

Low

SKILL.md:12

Hardcoded URL

- CloudBase main entry: `https://cnb.cool/tencent/cloud/cloudbase/cloudbase-skills/-/git/raw/main/sk

This is a hardcoded remote documentation URL used as a fallback source. It does not exfiltrate data, but it can make agents consume instructions outside the audited package.

Low

SKILL.md:13

Hardcoded URL

- Current skill raw source: `https://cnb.cool/tencent/cloud/cloudbase/cloudbase-skills/-/git/raw/mai

This is a hardcoded remote documentation URL used as a fallback source. It does not exfiltrate data, but it can make agents consume instructions outside the audited package.

Low

SKILL.md:29

Hardcoded URL

- Web application integration -> `../relational-database-web/SKILL.md` (standalone fallback: `https:

This is a hardcoded remote documentation URL used as a fallback source. It does not exfiltrate data, but it can make agents consume instructions outside the audited package.

Low

SKILL.md:30

Hardcoded URL

- Raw HTTP database access -> `../http-api/SKILL.md` (standalone fallback: `https://cnb.cool/tencent

This is a hardcoded remote documentation URL used as a fallback source. It does not exfiltrate data, but it can make agents consume instructions outside the audited package.

Risk Factors

⚙️ External commands (50)

🌐 Network access (4)

SKILL.md:12 SKILL.md:13 SKILL.md:29 SKILL.md:30

📁 Filesystem access (2)

SKILL.md:29 SKILL.md:30

Audited by: codex

← Back to Skill