Meeting Quality Scorer

• Outcome: Pass
• Latest audit: Jun 24, 2026 · 4:36 AM UTC
• Version: 1.0.0

Overview

This skill does what it advertises, but users should treat meeting transcripts as sensitive because they are sent to the configured LLM endpoint.

Install only if you are comfortable sending meeting transcript contents to the LLM provider you configure. Prefer a trusted local or approved provider for confidential, legal, HR, or regulated meetings, and pin/update dependencies before production use.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (10)

Vague Triggers

Medium

• Confidence: 93% confidence
• Finding: The trigger keywords include broad natural-language phrases such as 'score this meeting' and 'rate the meeting' that are likely to appear in ordinary conversation or unrelated contexts. This can cause unintended skill activation, leading to accidental transcript processing, file access, or transmission of sensitive meeting content to a configured LLM endpoint without the user's explicit intent.

Missing User Warnings

Medium

• Confidence: 86% confidence
• Finding: The function sends the full meeting transcript to an arbitrary OpenAI-compatible LLM endpoint via llm.chat() with no consent, warning, minimization, or visible trust boundary in this file. Meeting transcripts often contain sensitive business or personal data, so silent transmission to third-party or self-configured providers can cause unintended data disclosure and compliance/privacy issues.

Missing User Warnings

Medium

• Confidence: 86% confidence
• Finding: This second analysis path independently sends the transcript to the LLM again, increasing exposure of the same sensitive meeting content without any explicit warning in this file. Repeated full-transcript submission broadens privacy risk, especially because the skill is designed to work with many OpenAI-compatible endpoints of varying trustworthiness.

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: The script sends full transcript content to a configurable LLM endpoint, which may be a remote network service, without any explicit user warning or confirmation about data transmission. Meeting transcripts commonly contain sensitive business, personal, or regulated information, so silent exfiltration to third-party infrastructure creates a real confidentiality risk in this skill context.

Unpinned Dependencies

Low

• Category: Supply Chain

Content

openai>=1.30.0
pyyaml>=6.0
jinja2>=3.1.0
pytest>=8.0.0

• Confidence: 94% confidence
• Finding: openai>=1.30.0

Unpinned Dependencies

Low

• Category: Supply Chain

Content

openai>=1.30.0
pyyaml>=6.0
jinja2>=3.1.0
pytest>=8.0.0

• Confidence: 94% confidence
• Finding: pyyaml>=6.0

Unpinned Dependencies

Low

• Category: Supply Chain

Content

openai>=1.30.0
pyyaml>=6.0
jinja2>=3.1.0
pytest>=8.0.0

• Confidence: 97% confidence
• Finding: jinja2>=3.1.0

Unpinned Dependencies

Low

• Category: Supply Chain

Content

openai>=1.30.0
pyyaml>=6.0
jinja2>=3.1.0
pytest>=8.0.0

• Confidence: 90% confidence
• Finding: pytest>=8.0.0

Known Vulnerable Dependency: jinja2==3.1.0 — 5 advisory(ies): CVE-2025-27516 (Jinja2 vulnerable to sandbox breakout through attr filter selecting format metho); CVE-2024-56201 (Jinja has a sandbox breakout through malicious filenames); CVE-2024-22195 (Jinja vulnerable to HTML attribute injection when passing user input as keys to ) +2 more

Low

• Category: Supply Chain
• Confidence: 91% confidence
• Finding: jinja2==3.1.0

Known Vulnerable Dependency: pytest==8.0.0 — 1 advisory(ies): CVE-2025-71176 (pytest has vulnerable tmpdir handling)

Low

• Category: Supply Chain
• Confidence: 78% confidence
• Finding: pytest==8.0.0

VirusTotal

64/64 vendors flagged this skill as clean.