Back to skill

Security audit

GitLab

Security checks for vulnerabilities and agentic risk

Overview

GitLab changes require approval, but automatic setup executes unverified downloads and may change other agents’ installed skills without disclosure.

Review setup before enabling this skill. Install a verified CLI yourself, review its automatic skill synchronization, and use narrowly scoped GitLab permissions. Keep approval requirements for every write and deletion. The complete bundled Markdown was inspected; remote installer, CLI, and server implementations were not verified.

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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • 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
Findings (5)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The supplied command accesses an Oomol installation endpoint and executes its response, rather than performing a GitLab operation. Nothing in the supplied chunk establishes this installer as a supporting dependency for GitLab functionality. The remote script's contents and subsequent effects are not visible.

Content

No source excerpt is available for this finding.

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 55)May include surrounding context.

md
- `list_merge_request_notes` — List the notes (comments and system activity) on a GitLab merge request, to read a thread before answering it.
- `list_project_issues` — List issues for a GitLab project with common state, label, assignee, and search filters.
- `list_project_merge_requests` — List merge requests for a GitLab project with state, branch, search, and pagination filters.
- `list_projects` — List GitLab projects visible to the authenticated personal access token, with optional search and membership filters.
- `merge_merge_request` — Accept and merge a GitLab merge request. [write]
- `update_merge_request` — Update a GitLab merge request. [write]
- `update_project` — Update basic settings for an existing GitLab project. [write]

External Script Fetching

High
Category
Supply Chain
Confidence
98% confidence
Finding

Piping an unpinned HTTPS installer directly into bash executes whatever code the remote endpoint currently supplies, without integrity verification or an opportunity to inspect the complete script first. HTTPS protects transport but not against a compromised publisher or distribution pipeline, making this a supply-chain execution risk despite its legitimate setup purpose.

Content

Scanner excerpt · SKILL.md (reported line 74)May include surrounding context.

Fallback setup instructs execution of downloaded installer scripts without a system-integrity warning.

Medium
Category
Output Handling
Confidence
94% confidence
Finding

The fallback directs an agent to execute downloaded shell or PowerShell code without a separate installation approval or integrity-verification step. Although dependency setup is legitimate and conditional, compromise of the installer host or distribution pipeline could execute arbitrary code with the agent's local permissions; a warning alone would not prevent this.

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 95)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