Security audit

azure-ai-voicelive-py-free

Security checks for vulnerabilities and agentic risk

Overview

The skill appears to be an Azure VoiceLive helper, but it asks for broad read/write/command permissions and contains unclear generic file and command-execution guidance beyond the voice workflow.

Review before installing. Use this only in a constrained environment, grant microphone and Azure credentials only for intentional voice sessions, pin dependencies, and avoid granting general write or command execution unless the publisher narrows and documents exactly what commands and files are allowed.

A.I.G

Vulnerability Patterns

Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Insecure DependenciesIntroduces malicious components through unsafe dependency sources
Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review

Findings (2)

T08 · Insecure Dependencies

Warning

Location: SKILL.md:87
Finding: Unpinned Third-Party Dependencies Create Supply-Chain Risk
Content: View full analysis
Remediation: View remediation
aiohttp== ``` 2. Place direct and transitive dependencies in a locked requirements file. 3. Generate and verify cryptographic hashes, then install with: ```bash pip install --require-hashes -r requirements.txt ``` 4. Document the approved package index and disable untrusted extra indexes. 5. Scan pinned packages for known vulnerabilities before publishing the Skill. 6. Review and update dependencies through a controlled release process rather than automatically accepting the newest release. 7. Prefer a prebuilt, isolated runtime containing reviewed dependencies so the Skill does not need to install packages during invocation. ]]>

T05 · Unauthorized Access and Privilege Escalation

Note

Location: SKILL.md:26
Finding: Declared Tool Permissions Exceed the Demonstrated Voice Workflow
Content: View full analysis
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (5)

Missing User Warnings

High

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The skill describes microphone audio upload and transcript generation but does not clearly warn that audio and derived text may be sent to an external Azure service for processing. This omission can cause unintentional disclosure of sensitive spoken content, credentials, personal data, or regulated information to a third party without informed user consent.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 87% confidence
Finding: The display name, summary, and the entire user-facing documentation are presented only in Chinese, with no indication that users may choose another language. This can constitute a language/locale policy violation when a skill imposes a specific language without opt-in or documented regional justification.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 89% confidence
Finding: The invocation guidance tells users to call the skill with necessary input parameters but does not define clear trigger boundaries, permitted tasks, or disallowed uses. For a skill with read/exec/write tools listed, this ambiguity can encourage overly broad invocation and make it easier for downstream agents to apply the skill outside its intended narrow voice-processing context.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The document asserts that command execution is restricted to whitelist commands in a sandbox, but provides no implementation or enforceable policy showing that such controls exist. This creates a dangerous false sense of safety around exec-capable behavior, which could lead users or orchestrators to approve command execution that is not actually constrained.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The skill is presented as a narrowly scoped Azure VoiceLive wrapper, but later documentation expands its apparent capabilities to generic file handling and command execution. This scope inflation can mislead an agent or user into granting broader trust and invoking the skill in contexts unrelated to voice streaming, increasing the risk of unsafe file or shell operations under false assumptions about the skill’s purpose.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.