ClawHub
Back to skill

Security audit

Databricks

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly coherent for Databricks administration, but its setup instructions include executing a remote installer script without verification.

Review the oo CLI installer before running it, prefer a verified package or checksum-backed install path if available, and only allow write or destructive Databricks actions after checking the exact target and payload.

A.I.G

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

SkillSpector

Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 51)May include surrounding context.

md
- `delete_secret_scope` — Delete a Databricks secret scope by scope name. [destructive]
- `edit_cluster` — Edit an existing Databricks cluster by cluster ID using a raw official clusters edit payload. [write]
- `get_cluster` — Get one Databricks cluster by cluster ID.
- `get_current_user` — Get the current Databricks workspace principal profile for the connected personal access token.
- `get_job` — Get one Databricks job, including its current settings and optional continuation page token.
- `get_run_by_id` — Get one Databricks job run by run ID.
- `get_run_output` — Get the output payload for one Databricks run, including notebook output or error details when available.

External Script Fetching

High
Category
Supply Chain
Confidence
98% confidence
Finding

The skill includes a one-line remote install pattern that downloads and immediately executes a script via 'curl ... | bash'. This is dangerous because it gives a remote server direct code execution on the host without integrity verification, and a compromised server, MITM in a weak environment, or malicious script update could lead to full system compromise.

Content

Scanner excerpt · SKILL.md (reported line 90)May include surrounding context.

Static analysis

No suspicious patterns detected.