Apple Silicon Ai

• Outcome: Pass
• Latest audit: Sep 16, 2026 · 4:08 AM UTC
• Version: 1.0.3

Overview

The skill is a coherent local Apple Silicon AI fleet guide with disclosed local services and state, though users should treat the unpinned pip install as normal supply-chain risk.

Before installing, consider using a virtual environment and pinning or verifying the `ollama-herd` package version, especially if deploying across multiple Macs. Expect a local HTTP service on port 11435, local-network node discovery, and local logs/metrics under `~/.fleet-manager`.

A.I.G

Vulnerability Patterns

• Insecure Dependencies: Introduces malicious components through unsafe dependency sources
• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally

Findings (1)

T08 · Insecure Dependencies

Warning

• Location: SKILL.md:30

Finding

Unpinned Third-Party Package Installation

Content

View full analysis

Vulnerability Details

File Location: SKILL.md, lines 30–34
Vulnerability Type: Unpinned and unverified third-party dependency
Risk Level: Medium

Vulnerable Code

markdown

### 1. Install on every Apple Silicon Mac

```bash
pip install ollama-herd    # Apple Silicon optimized inference router

text

### Technical Analysis

The installation instructions direct users to install `ollama-herd` from pip's configured package index without pinning a version or verifying an artifact hash, signature, source commit, or release provenance. Although the Skill identifies a GitHub homepage, the audited project contains no lockfile or verification data that cryptographically binds the installed package to that repository or to a reviewed release.

Python package installation may execute package-controlled build and installation logic. The installed application is subsequently launched through the `herd` and `herd-node` commands. Consequently, a compromised package release, publishing account, package-index artifact, transitive dependency, or unexpected future version could introduce code that was not part of this audit.

### Attack Path

1. An attacker compromises the package publisher, package-index artifact, release pipeline, or an unpinned transitive dependency.
2. The attacker publishes a malicious or modified package version under the expected package name.
3. A user follows the documented `pip install ollama-herd` command, which resolves the current package version from the configured index.
4. Pip downloads and installs attacker-controlled package content and may execute package-controlled build logic.
5. The user starts `herd` or `herd-node`, causing the modified runtime code to execute.
6. The malicious code operates with the privileges and network access of the installing or executing user.

This is a supply-chain exposure rather than evidence that the named package is currently malicious.

### Impa
...[truncated 723 chars]

Remediation

View remediation

Remediation Suggestions

1. Pin the dependency to a specifically reviewed version, using an exact version constraint rather than an unconstrained package name.
2. Commit a requirements or lock file containing cryptographic hashes, and install with pip's --require-hashes option.
3. Document the trusted package index explicitly and disable unintended supplemental indexes to reduce dependency-confusion risk.
4. Publish and verify signed releases or attestations that bind package artifacts to the reviewed source commit and build process.
5. Audit and lock transitive dependencies in addition to the top-level package.
6. Install the package inside a dedicated virtual environment under a non-privileged account; do not use elevated installation unless strictly necessary.
7. Include the reviewed package source or reproducible-build metadata in the audit artifact so documented network, file-system, and download claims can be independently verified.
8. Re-audit each package update before changing the pinned version or hashes.

SkillSpector

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (2)

External Transmission

Medium

• Category: Data Exfiltration
• Confidence: 60% 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 67)May include surrounding context.

OpenAI-compatible API (Apple Silicon backend)

bash

curl http://localhost:11435/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "llama3.3:70b",

Natural-Language Policy Violations

Low

• Category: Not specified by scanner
• Confidence: 93% confidence

Finding

The manifest description mixes English, Chinese, and Spanish in a mandatory skill description without indicating that language selection is optional or user-configurable. The policy for natural-language violations applies to all file types and flags locale or language constraints when they are imposed without opt-in.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.