Back to skill

Security audit

diagnose-ollama-calls

Security checks for vulnerabilities and agentic risk

Overview

The skill’s local troubleshooting behavior fits its purpose, with caveats about downloaded software and retained diagnostic records.

Use a reviewed, pinned package version consistently for installation and MCP startup. Route only intended calls through the observer, keep it local, and review recording settings before handling sensitive prompts. The instructions were fully inspected; they do not establish guarantees about the separately downloaded software.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

MCP server referenced without pinned version: 'npx llmxray'.

Medium
Category
MCP Rug Pull
Confidence
97% confidence
Finding

The unversioned npx llmxray command can download and execute a changing package version with the user's privileges. A compromised release could access local files or capture the model requests routed through the proxy; claims that everything stays local do not establish that downloaded code is trustworthy.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% 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 70)May include surrounding context.

md
| `estimate_my_prompt` | A prompt is growing, or you are about to raise `num_ctx` or switch model. | Fit and headroom against the loaded context, VRAM fit, likely cache reuse. |
| `read_llmxray_status` | Another tool says nothing was seen. | The observe port, the Ollama behind it, turns kept. |

All seven take no secrets and change nothing. Call them without asking the user first.

## Procedure

MCP server referenced without pinned version: 'npx llmxray'.

Medium
Category
MCP Rug Pull
Confidence
97% confidence
Finding

The troubleshooting instructions repeat the unpinned package-execution command, preserving the same supply-chain exposure as the initial setup. If an untrusted release is resolved, it runs locally with the user's privileges and is positioned to observe potentially sensitive model traffic.

Content

No source excerpt is available for this finding.