Back to skill

Security audit

Qmd2

Security checks for vulnerabilities and agentic risk

Overview

This skill is a small, coherent local search helper, with a disclosed local index and a notable but not disqualifying unpinned GitHub install dependency.

Before installing, consider that the qmd dependency is pulled from an unpinned GitHub source and that indexed file contents may be stored locally under ~/.cache/qmd. Index only directories you intend to make searchable, and prefer a pinned or reviewed qmd revision if you need stronger supply-chain control.

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:4
Finding

Unpinned Mutable GitHub Dependency

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, line 4
Vulnerability Type: Unpinned third-party dependency from a mutable remote repository
Risk Level: Medium

Vulnerable Code Snippet:

yaml
metadata: {"clawdbot":{"emoji":"📝","requires":{"bins":["qmd"]},"install":[{"id":"node","kind":"node","package":"https://github.com/tobi/qmd","bins":["qmd"],"label":"Install qmd (node)"}]}}

Technical Analysis

The installation metadata references https://github.com/tobi/qmd without pinning an immutable commit SHA, versioned release artifact, or integrity checksum. Consequently, the code retrieved during installation may change after this Skill has been reviewed.

This creates a supply-chain trust boundary: installation and runtime security depend on the continuing integrity of the remote repository, its maintainers, and its transitive dependencies. If that source is compromised or modified unexpectedly, a later installation could execute code different from the code originally audited. The package implementation is not included in this project, so its installation scripts, runtime operations, and transitive dependencies could not be verified during this audit.

Attack Path

  1. An attacker compromises the referenced repository, a maintainer account, the package's release process, or a dependency used by the remote project.
  2. The attacker adds malicious Node.js installation or runtime code to the mutable upstream source.
  3. A user installs the Skill dependency using the declared GitHub URL.
  4. The installation mechanism retrieves the altered upstream content because no immutable revision or integrity value is specified.
  5. The malicious code executes with the privileges of the account performing installation or invoking qmd.
  6. Within those privileges, the payload could access files readable by that account, including files selected for indexing and generated index data, or communicate data to an attack ...[truncated 907 chars]
Remediation
View remediation

Remediation Suggestions

  1. Replace the mutable repository reference with an immutable, reviewed commit SHA or a versioned release artifact.
  2. Verify the downloaded artifact using a cryptographic checksum or supported package-manager integrity metadata.
  3. Prefer a trusted registry package only when its package name, publisher identity, provenance, and release history have been verified.
  4. Lock all transitive dependencies with an auditable lockfile and use reproducible or frozen dependency installation.
  5. Review upstream installation scripts and runtime code before approving a pinned revision.
  6. Enable automated dependency and provenance monitoring, but require review before updating the pinned revision.
  7. Run the tool with least privilege and restrict its file and network access to the minimum required for local indexing and configured Ollama communication.
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