Back to skill

Security audit

Asb

Security checks for vulnerabilities and agentic risk

Overview

The skill is a disclosed ASB/MCP integration, but it asks users to install and run unpinned external code with broad orchestration, LLM, and message-sending capabilities that are not clearly bounded.

Review this before installing if you plan to connect it to real services or sensitive data. Prefer a pinned, reviewed ASB release, run it with least privilege, and treat message-send, orchestration, and LLM-provider tools as capable of sending data externally or triggering downstream operations.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:16
Finding

Unpinned Third-Party Package and Repository Installation

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 16–18 and 124–132
Vulnerability Type: T08: Insecure Dependencies
Risk Level: Medium

The Skill directs users to install and execute externally maintained code without pinning an exact package version, immutable Git commit, artifact checksum, or cryptographic signature.

Affected code at lines 16–18:

yaml
install:
  - kind: uv
    package: agent-service-bus

Affected code at lines 124–132:

bash
# Clone the repository
git clone https://github.com/qiyueqiu/Agent-Service-Bus.git
cd Agent-Service-Bus

# Install with uv
uv sync

Technical Analysis

The metadata installation retrieves agent-service-bus without an exact version constraint. The documented alternative clones the mutable default branch of the external Git repository and then uses uv sync to resolve and install its dependencies.

Neither path establishes an immutable relationship between the reviewed Skill and the code ultimately executed. The project contains only SKILL.md; therefore, the implementation of the installed ASB CLI and MCP server, as well as its transitive dependencies, cannot be verified within the audited artifact.

Because package releases, repository branches, and dependency-resolution results can change after review, a future compromised or malicious revision could be installed under the same instructions. This is a supply-chain weakness rather than evidence that the currently referenced project is malicious.

Attack Path

  1. An attacker compromises the package registry entry, upstream repository, maintainer account, release process, or an unpinned transitive dependency.
  2. The attacker publishes a malicious package release or modifies the repository's mutable default branch.
  3. A user installs the Skill dependency through the unversioned agent-service-bus declaration, or follows the documented git clone and uv sync procedu ...[truncated 1054 chars]
Remediation
View remediation

Remediation Suggestions

  1. Pin agent-service-bus to an exact, reviewed package version rather than allowing the latest available release.
  2. For Git-based installation, check out a full immutable commit hash instead of relying on the repository's mutable default branch.
  3. Publish and verify cryptographic checksums or trusted signatures for downloaded artifacts.
  4. Commit and enforce a dependency lockfile that records exact transitive dependency versions and integrity information.
  5. Ensure the metadata installation and manual installation instructions resolve to the same audited release.
  6. Review the ASB implementation and its complete dependency tree before exposing it as an MCP server.
  7. Run the MCP server with least privilege, restricted filesystem access, a sanitized environment, and limited network access.
  8. Add automated dependency provenance, vulnerability, and integrity checks to the release process.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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
Findings (1)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The skill advertises message-send and orchestration capabilities but does not warn users that these actions may transmit data to external systems or trigger downstream operations. In an agent-integrated environment, this can lead to unintended disclosure of sensitive data or execution of side effects when a user assumes the tools are local or read-only.

Content

No source excerpt is available for this finding.