Back to skill

Security audit

Seedance 2.5 Reference to Video — 1080p Reference-Guided Video on RunComfy

Security checks for vulnerabilities and agentic risk

Overview

The skill appears purpose-aligned, but it needs review because it uses an unpinned third-party CLI and has ambiguous routing and input guidance for a paid remote video service.

Before installing, confirm you are comfortable sending prompts and user-selected reference media URLs to RunComfy and with the stated per-second billing. Prefer a pinned, verified `@runcomfy/cli` version, avoid running `npx -y` with broad privileges, and use only reference URLs explicitly chosen for the generation.

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:63
Finding

Unpinned Third-Party CLI Installation and Execution

Content
View full analysis
Remediation
View remediation
npx --yes @runcomfy/cli@ ``` 2. Document the expected package version, official npm registry, publisher identity, and package integrity hash so users can verify what will execute. 3. Prefer a committed lockfile and `npm ci` for managed deployments rather than dynamically resolving the latest release. 4. Disable npm lifecycle scripts where compatible with the package: ```bash npm install --global --ignore-scripts @runcomfy/cli@ ``` Confirm first that the legitimate CLI does not require installation scripts. 5. Recommend executing the CLI under a non-privileged account or within a restricted container. Grant access only to the intended output directory and the RunComfy credential required for the request. 6. Avoid exposing unrelated secrets to the CLI process. Supply only `RUNCOMFY_TOKEN`, and use a narrowly scoped or short-lived token if RunComfy supports one. 7. Establish an update-review process: verify new CLI releases and their dependency changes before updating the pinned version. ]]>
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (5)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger phrases include broad terms like reference to video, seedance 1080p, and generic asks to generate video from reference images. Overbroad activation can cause the wrong skill to run for ordinary user requests, leading to unintended submission of prompts and reference URLs to a paid remote service.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The top-level description and invocation guidance do not clearly foreground that prompts and user-supplied reference URLs are sent to a third-party service. Although the later Security & Privacy section discloses this, users or routing systems may rely on the earlier description and trigger metadata, creating a meaningful consent and privacy transparency gap.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The document states in one section that videos is optional and can be omitted, but later lists a reference video as mandatory. Conflicting guidance about required inputs can cause an agent or user to submit unintended reference media, potentially increasing data disclosure to the third-party API and unexpectedly raising billable usage.

Content

No source excerpt is available for this finding.