πŸ‘„ Lipsync β€” Pro Pack on RunComfy

Security checks across static analysis, malware telemetry, and agentic risk

Overview

The skill appears benign: it openly uses RunComfy’s CLI and account token to generate lip-sync videos, with normal third-party media sharing and consent considerations.

Before installing, make sure you trust the RunComfy CLI package and are comfortable sending the selected audio/video media to RunComfy. Use a properly scoped token, keep it private, and only create lip-sync content for people and media you have permission to use.

Static analysis

No static analysis findings were reported for this release.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Risk analysis

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

#
ASI03: Identity and Privilege Abuse
Low
What this means

Using the skill may consume RunComfy account resources and gives the CLI access to submit jobs on the user’s behalf.

Why it was flagged

The skill uses a RunComfy account token/login so CLI actions run under the user’s RunComfy account. This is expected for the service integration and no credential logging or unrelated use is shown.

Skill content
runcomfy login              # or in CI: export RUNCOMFY_TOKEN=<token>
Recommendation

Use a RunComfy token with appropriate scope, avoid sharing tokens, and revoke or rotate the token if it is no longer needed.

#
ASI04: Agentic Supply Chain Vulnerabilities
Low
What this means

Installing a global CLI package can affect the local environment if the package source is compromised or unexpected.

Why it was flagged

The skill documents installing and running an external npm CLI package. This is central to the stated purpose, but it still relies on the npm package’s provenance and updates.

Skill content
npm i -g @runcomfy/cli      # or:  npx -y @runcomfy/cli --version
Recommendation

Install the CLI from the official RunComfy documentation or a trusted package source, and review package details before using global installation.

#
ASI07: Insecure Inter-Agent Communication
Low
What this means

Media URLs or referenced media content may be processed by RunComfy and its selected model providers.

Why it was flagged

The workflow sends user-provided video and audio URLs to RunComfy model endpoints and writes outputs locally. The provider and data flow are disclosed and purpose-aligned.

Skill content
runcomfy run <vendor>/<model> \
  --input '{"video_url": "...", "audio_url": "..."}' \
  --output-dir ./out
Recommendation

Only provide media you have rights to share with RunComfy, and review RunComfy’s privacy, retention, and provider-routing terms for sensitive content.

#
ASI09: Human-Agent Trust Exploitation
Low
What this means

Improper use could create harmful or deceptive videos involving real people.

Why it was flagged

Lip-sync generation can be used to create misleading synthetic media, but the artifact explicitly discloses the dual-use risk and gives refusal guidance.

Skill content
Driving a real person's mouth from a separate audio track is dual-use. Refuse user requests that target real public figures without consent, or that aim at defamatory or sexually explicit synthetic media.
Recommendation

Use the skill only with consent and avoid public-figure impersonation, defamatory content, sexual synthetic media, or other deceptive uses.