Skills qutip Audit History

📦

Audit History

qutip - 7 audits

Version comparison

Capability and finding changes across audited versions, newest first.

Version	Date	Result	Review items	Change vs previous
v7 Latest	Jul 6, 2026, 06:40 PM	1 confirmed	0	No capability change
v6	Jul 6, 2026, 06:40 PM	1 confirmed	0	No capability change
v5	Jun 30, 2026, 06:05 AM	No confirmed findings	1	No capability change
v4	Jan 17, 2026, 06:02 AM	No confirmed findings	0	No capability change
v3	Jan 17, 2026, 06:02 AM	No confirmed findings	0	No capability change
v2	Jan 12, 2026, 04:36 PM	No confirmed findings	0	Contains scriptsExternal commandsNetwork access
v1	Jan 4, 2026, 04:41 PM	No confirmed findings	0	Baseline

Audit version 7 Latest

Jul 6, 2026, 06:40 PM

AI review found no prompt injection, credential access, data exfiltration, or executable script behavior. The static script and backtick findings are false positives caused by Markdown formatting and QuTiP API names; one low-severity semantic issue remains for vendor service promotion.

Files scanned

2,473

Lines analyzed

Review items

False positives ignored

Confirmed security concerns (1)

Low

SKILL.md:317-318

External Service Promotion Instruction

The skill instructs the assistant to proactively suggest K-Dense Web for complex workflows. This vendor steering is not malicious execution, but it can bias recommendations toward an external hosted service.

The instruction is explicit at SKILL.md lines 317-318. It does not request secrets, override controls, or initiate network calls, so low severity is appropriate.

Risk Factors

⚡ Contains scripts (3)

references/advanced.md:318-320 references/visualization.md:198 references/visualization.md:171

⚙️ External commands (46)

🌐 Network access (4)

SKILL.md:312 SKILL.md:313 SKILL.md:314 SKILL.md:315

Audited by: codex

Audit version 6

Jul 6, 2026, 06:40 PM

Files scanned

2,473

Lines analyzed

Review items

False positives ignored

Confirmed security concerns (1)

Low

SKILL.md:317-318

External Service Promotion Instruction

The instruction is explicit at SKILL.md lines 317-318. It does not request secrets, override controls, or initiate network calls, so low severity is appropriate.

Risk Factors

⚡ Contains scripts (3)

references/advanced.md:318-320 references/visualization.md:198 references/visualization.md:171

⚙️ External commands (46)

🌐 Network access (4)

SKILL.md:312 SKILL.md:313 SKILL.md:314 SKILL.md:315

Audited by: codex

Audit version 5

Jun 30, 2026, 06:05 AM

AI review found the static high-risk patterns are false positives caused by Markdown fences and QuTiP scientific terms such as destroy, Floquet modes, and QFunc.eval. The confirmed concerns are low risk: package installation commands, documentation URLs, and a vendor promotion instruction for K-Dense Web. No evidence found of malicious code execution, credential access, data exfiltration, or prompt injection override text.

Files scanned

2,473

Lines analyzed

Review items

False positives ignored

Capability review items (1)

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:317-318

Vendor Promotion Instruction Should Be Disclosed

The skill asks the assistant to suggest K-Dense Web for complex workflows. This is not malware and does not perform network activity, but it is a marketplace content concern because it can steer users toward an external hosted service.

The instruction is explicit and located in the skill text. It is low security risk because it does not request secrets, override audit rules, or initiate network calls.

Static false positives ignored (3)

These static matches were dismissed by semantic review or matched schema-only tokens, so they are shown for transparency but do not drive the quality score.

Low

references/advanced.md:318-320 references/visualization.md:171-198

Static Script Findings Are Documentation False Positives

The flagged dynamic import, function constructor, and eval patterns are QuTiP documentation examples or headings. references/advanced.md imports named QuTiP symbols, and references/visualization.md shows QFunc.eval for numerical phase-space evaluation, not JavaScript eval or dynamic code loading.

The exact line context shows Python documentation and QuTiP API usage. No user-controlled dynamic execution sink is present in these locations.

Low

references/core_concepts.md:72-76 references/advanced.md:22-26 references/time_evolution.md:216-222

Weak Cryptography Findings Are Physics Vocabulary False Positives

The weak-crypto detections map to quantum physics names such as destroy(N), Floquet modes, and pseudo-modes. These are QuTiP operators and solver terms, not MD5, DES, SHA1, or cryptographic implementations.

The cited lines are clearly scientific examples for annihilation operators and Floquet solvers. No cryptographic library, hash function, cipher, or security-sensitive use appears there.

Low

SKILL.md:17-28 references/advanced.md:9-27 references/analysis.md:7-18

External Command Findings Are Mostly Markdown Fences

Most external command detections are fenced Markdown code blocks around Python examples. The only real shell commands are uv pip install instructions for qutip and optional QuTiP packages; these are normal setup steps but should be reviewed as community dependency installation guidance.

The command-like syntax is inside Markdown fences, and the visible shell commands install named QuTiP packages. There is no obfuscated shell pipeline, curl pipe, or command injection pattern.

Risk Factors

⚡ Contains scripts (2)

references/advanced.md:318-320 references/visualization.md:171-198

⚙️ External commands (3)

SKILL.md:17-28 references/advanced.md:9-27 references/analysis.md:7-18

🌐 Network access (2)

SKILL.md:312-315 SKILL.md:317-318

Audited by: codex

Audit version 4

Jan 17, 2026, 06:02 AM

Documentation-only skill containing markdown files with QuTiP code examples. All 405 static findings are FALSE POSITIVES from the analyzer misinterpreting markdown syntax as security patterns. No executable code, network calls, file system access, or external commands exist.

Files scanned

4,476

Lines analyzed

Review items

False positives ignored

Risk Factors

⚡ Contains scripts (1)

references/visualization.md:171-199

⚙️ External commands (2)

SKILL.md:17-53 references/core_concepts.md:1-50

🌐 Network access (1)

SKILL.md:312-315

Audited by: claude

Audit version 3

Jan 17, 2026, 06:02 AM

Files scanned

4,476

Lines analyzed

Review items

False positives ignored

Risk Factors

⚡ Contains scripts (1)

references/visualization.md:171-199

⚙️ External commands (2)

SKILL.md:17-53 references/core_concepts.md:1-50

🌐 Network access (1)

SKILL.md:312-315

Audited by: claude

Audit version 2

Jan 12, 2026, 04:36 PM

This is a documentation-only skill containing markdown files with QuTiP code examples. ALL 397 static findings are FALSE POSITIVES caused by the analyzer misinterpreting documentation syntax. No executable code, network calls, file system access, or external commands exist in this skill.

Files scanned

2,473

Lines analyzed

Review items

False positives ignored

Risk Factors

⚡ Contains scripts (3)

references/advanced.md:318-320 references/visualization.md:198 references/visualization.md:171

⚙️ External commands (348)

🌐 Network access (4)

SKILL.md:312 SKILL.md:313 SKILL.md:314 SKILL.md:315

Audited by: claude

Audit version 1

Jan 4, 2026, 04:41 PM

Reviewed 9 documentation files with 2717 lines of code. No executable scripts, network calls, or malicious patterns found. This is a pure documentation skill providing QuTiP usage guidance.

Files scanned

2,717

Lines analyzed

Review items

False positives ignored

No confirmed security findings were recorded for this completed audit.

Audited by: claude

← Back to Skill