Back to skill

Security audit

百积木 Bundle 开发

Security checks for vulnerabilities and agentic risk

Overview

This skill is a coherent Baijimu bundle-development helper, with expected CLI installation and user-directed publishing/runtime actions but no hidden or destructive behavior in the artifact.

Install this only if you intend to use Baijimu Bundle workflows and trust the @baijimu/cli npm package source. Use a scoped Baijimu account/workspace, verify the CLI version and docs before acting, and review publish/install/uninstall actions before allowing them to run.

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:12
Finding

Unpinned Executable npm Dependency

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 12–15
Vulnerability Type: Unpinned third-party executable dependency
Risk Level: Medium

yaml
install:
  - kind: node
    package: "@baijimu/cli"
    bins: [baijimu]

Technical Analysis

The Skill declares installation of the executable npm package @baijimu/cli without specifying an exact version, lockfile, or integrity digest. Package resolution can therefore select mutable registry content that was not included in or verified by this audit. The package implementation and any installation lifecycle scripts are absent from the audited project.

This creates a supply-chain trust boundary: compromise of the package, its maintainer account, registry resolution, or a transitive dependency could cause attacker-controlled code to execute when the dependency is installed or invoked.

Attack Path

  1. An attacker compromises the npm package, a maintainer account, registry delivery, or a dependency in its supply chain.
  2. The attacker publishes or causes resolution to a malicious release.
  3. A user or Agent loads the Skill and installs the dependency using the unversioned package declaration.
  4. npm resolves the attacker-controlled version.
  5. Malicious lifecycle-script or CLI code executes with the privileges of the installing process.
  6. The code may access local files and credentials or interfere with authenticated Baijimu publication and Runtime operations available to that process.

Impact Assessment

Successful exploitation could obtain code execution under the installing user's privileges. The accessible scope may include local project data, environment variables, CLI credentials, authenticated Baijimu workspaces, and Bundle publication or Runtime operations authorized to that identity. No malicious payload was present in the audited repository itself; exploitation depends on compromise or unsafe resolution of the external dependency.

Remediation
View remediation

Remediation Suggestions

  • Pin @baijimu/cli to a reviewed, exact version rather than using an unconstrained package reference.
  • Record and verify the package integrity digest through an appropriate lockfile or equivalent installation metadata.
  • Use an explicitly trusted npm registry and enforce registry provenance controls where supported.
  • Review the pinned package, its transitive dependency graph, and any npm lifecycle scripts before deployment.
  • Disable lifecycle scripts during installation where they are not required.
  • Run the CLI with least privilege and isolate it from credentials, files, and environments unrelated to the Bundle task.
  • Establish an update process that separately reviews and verifies each dependency version change before adoption.
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The user-facing skill content from the title onward is written entirely in Chinese, and the file does not indicate that Chinese is optional, user-selected, or required for a region-specific compliance reason. This creates a natural-language locale policy concern because it effectively imposes a single language on users without opt-in.

Content

No source excerpt is available for this finding.