Back to skill

Security audit

briq

Security checks for vulnerabilities and agentic risk

Overview

The skill is marked suspicious because a confirmed finding says its cleanup instructions could delete resources containing user data without warning; scan coverage was partial.

Review the cleanup guidance before installing: make sure you can protect any data stored in resources from unintended deletion. Also consider whether you trust the external Briq service and how it may handle information. The scan was partial, so the available evidence does not fully assess the skill.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 14)May include surrounding context.

md
# briq

briq runs services and sandboxes for agents, each on its own microVM, through one MCP server:
`https://api.briq.cloud/mcp` (Streamable HTTP, OAuth). Every answer about a briq shows its cost so
far and when it goes to sleep.

## Connect, if the briq tools are missing

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 22)May include surrounding context.

md
You have the tools when `briq_run`, `briq_exec` and `briq_quota` are available. If not:

1. Add the MCP server for all projects (user scope): name `briq`, URL
   `https://api.briq.cloud/mcp`. Your client's exact steps: https://briq.cloud/docs/setup/clients.md
2. Get the user signed in. They pick the team you work for and your budget; you never handle a
   password or a token.
   - On their computer: start your client's sign-in (`claude mcp login briq`,

The cleanup instruction directs the agent to destroy resources without warning about potential loss of their stored data.

Medium
Category
Output Handling
Confidence
94% confidence
Finding

The instruction to destroy resources once finished does not distinguish disposable sandboxes from databases or services holding persistent user data. An agent could interpret completion of its task as authorization to delete a data-bearing resource, even though the user expects its contents to remain available; the documented sleep option provides a safer alternative.

Content

No source excerpt is available for this finding.