Back to skill

Security audit

VideoMemory

Security checks for vulnerabilities and agentic risk

Overview

The skill is a narrowly scoped VideoMemory installer/launcher, with a normal third-party package installation risk users should understand.

Install this only if you trust the @clamepending/videomemory package and are comfortable with it adding OpenClaw bridge files and starting a local VideoMemory service. Review the --safe --explain onboarding plan before allowing setup, and run it in a constrained environment if you need stronger supply-chain isolation.

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
Unreviewed Third-Party Package Installation and Execution## Vulnerability Details **File Location**: `SKILL.md`, lines 16–39 **Vulnerability Type**: Third-party supply-chain risk caused by automatic package installation and execution **Risk Level**: Medium ### Vulnerable Code ```bash openclaw plugins install @clamepending/videomemory@0.1.7 ``` ```bash npx -y @clamepending/videomemory@0.1.7 onboard --safe --repo-ref v0.1.3 --explain ``` ```bash npx -y @clamepending/videomemory@0.1.7 onboard --safe --repo-ref v0.1.3 ``` ```bash npx -y @clamepending/videomemory@0.1.7 relaunch --repo-ref v0.1.3 ``` ### Technical Analysis The skill directs the agent to install and execute the external npm/OpenClaw package `@clamepending/videomemory@0.1.7`. The audited project does not contain the package implementation, integrity hashes, a lockfile, permission declarations, or other evidence sufficient to verify the downloaded code. The `npx -y` option downloads and executes package code without an interactive confirmation step. Installing the OpenClaw plugin also introduces executable third-party code into the environment. Although the package and repository reference are version-pinned, version pinning alone does not verify package integrity or protect against a compromised registry account, malicious package publication, registry tampering, or unsafe lifecycle scripts. The `--safe` argument does not independently enforce a security boundary because it is interpreted by the same unreviewed package. No malicious behavior is proven in the supplied artifact; the finding concerns the unsafe trust and execution model. ### Attack Path 1. An attacker compromises the package publisher, registry account, distribution infrastructure, or the referenced package artifact. 2. The user requests installation, onboarding, or relaunch through this skill. 3. The agent runs `openclaw plugins install` or `npx -y`, retrieving the external package. 4. The package or its installation lifecycle scripts ...[truncated 1199 chars]
Remediation
## Remediation Suggestions 1. Vendor and review the package implementation, including installation scripts, lifecycle hooks, onboarding behavior, relaunch behavior, and bridge-file modifications. 2. Verify publisher identity, registry provenance, release signatures, and the relationship between package version `0.1.7` and repository reference `v0.1.3`. 3. Pin the package using an immutable cryptographic integrity digest or a reviewed lockfile rather than relying only on a mutable registry name and version. 4. Avoid unattended `npx -y` execution. Require explicit user approval after presenting the package source, expected changes, requested permissions, and verified integrity information. 5. Disable package lifecycle scripts where compatible, or separately inspect and approve each required lifecycle operation. 6. Run installation and onboarding inside a sandbox or restricted account with minimal filesystem access, no unnecessary credentials, and tightly limited network permissions. 7. Restrict writable plugin, configuration, and startup locations to reduce the potential for persistent modification. 8. Record and verify the files, processes, ports, and network destinations created by onboarding. Alert on behavior outside the documented VideoMemory setup scope. 9. Treat `--safe` as an application option rather than a security control; enforce safety through operating-system permissions, network policy, and process isolation.
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

Static analysis

No suspicious patterns detected.