Back to skill

Security audit

Gladia

Security checks for vulnerabilities and agentic risk

Overview

The Gladia workflow includes approval safeguards, but setup directs the agent to run unverified remote installers without a clear approval step.

Review and install the CLI separately before enabling this skill, using a verified release where available. Consider its automatic skill synchronization; upstream documents OO_SKILLS_SYNC_DISABLED as a control. Expect requested audio and transcription data to pass through OOMOL and Gladia. Remote installer contents remain unverified; no malicious behavior was established.

Vulnerability Patterns
  • 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
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (4)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The supplied code executes a remotely hosted installer rather than performing Gladia operations. Nothing in the chunk establishes that this installation is a supporting dependency for Gladia. The remote script's contents are not supplied, so its further effects cannot be determined.

Content

No source excerpt is available for this finding.

External Script Fetching

High
Category
Supply Chain
Confidence
98% confidence
Finding

Piping a mutable HTTPS response directly into bash executes installer content without version pinning or independent integrity verification. A compromised distribution endpoint or malicious installer update could execute arbitrary code with the user's privileges; HTTPS alone does not protect against those supply-chain failures, and the conditional setup limits exposure but does not remove it.

Content

Scanner excerpt · SKILL.md (reported line 60)May include surrounding context.

The setup workflow executes downloaded installer scripts without warning the user about running remote code or modifying their system.

Medium
Category
Output Handling
Confidence
94% confidence
Finding

The fallback directs the agent to execute a downloaded shell or PowerShell installer without first obtaining explicit approval for local software installation. Although installing a required dependency is legitimate and conditional, this can run arbitrary remote code with the invoking user's privileges and modify the host beyond the requested Gladia operation.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
80% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 81)May include surrounding context.

md
- **`OOMOL_INSUFFICIENT_CREDIT`** — OOMOL billing stop. Recharge at `https://console.oomol.com/billing/token-recharge` before retrying.

- **HTTP 402 without `OOMOL_INSUFFICIENT_CREDIT`** — the third-party account itself has no active plan or credit. Report the provider error message as is; do not ask the user to recharge OOMOL.

## Resources