Back to skill

Security audit

Seedance 2.0 Pro — Pro Pack on RunComfy

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed RunComfy video-generation wrapper with a supply-chain caution around installing the external CLI globally.

Before installing, consider using a pinned or project-local @runcomfy/cli version, avoid sudo for npm installs, keep RUNCOMFY_TOKEN out of shell startup files, and only send prompts or media URLs you are comfortable sharing with RunComfy.

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:64
Finding

Unpinned Global Installation of an Externally Maintained CLI

Content
View full analysis
` instead of `runcomfy login`. ``` ### Technical Analysis The installation command retrieves the current release of `@runcomfy/cli` from the npm registry without pinning an exact version or verifying its integrity. It also installs the package globally, expanding its reach beyond an isolated project environment. The dependency's source code is not included in the audited project, so its installation scripts, runtime behavior, update process, and handling of `RUNCOMFY_TOKEN` cannot be independently verified from this artifact. A future malicious or compromised package release could execute npm lifecycle scripts during installation or malicious code when the CLI is invoked. This is a supply-chain weakness rather than evidence that the current RunComfy package is malicious. The network submission of prompts and media references to RunComfy is explicitly documented and is necessary for the declared hosted video-generation functionality. No hidden transmission to unrelated services was identified in `SKILL.md`. ### Attack Path 1. An attacker compromises the npm package, maintainer account, publishing credentials, or release pipeline for `@runcomfy/cli`. 2. The attacker publishes a malicious release under the legitimate package name. 3. A user follows the unpinned command `npm i -g @runcomfy/cli`, causing npm to retrieve that release. 4. Malicious lifecycle code may execute during installation, or the altered CLI may execute later when the Skill invokes `runcomfy`. 5. The malicious package runs with the privileges of the installing or invoking user. 6. It ma ...[truncated 1190 chars]
Remediation
View remediation
``` 2. Record and verify the expected package integrity hash and package provenance before installation. 3. Prefer a project-local installation governed by a committed lockfile instead of a global installation: ```bash npm install --save-exact @runcomfy/cli@ ``` 4. Invoke the locked local binary rather than resolving an arbitrary global executable. 5. Audit the selected CLI release, including npm lifecycle scripts, network destinations, credential handling, output download behavior, and cancellation logic. 6. Avoid installing npm packages with administrator privileges. Run the CLI in a restricted container or sandbox with access only to the required output directory and network endpoints. 7. Provide `RUNCOMFY_TOKEN` only to the generation process, avoid placing it in persistent shell configuration, and rotate it if dependency compromise is suspected. 8. Use a minimally privileged RunComfy token where supported, with usage limits and account monitoring. 9. Document the reviewed package version and establish an explicit update process that requires security review before changing it. ]]>
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