video-agent
Verdict safe, but 1 finding (1 medium), mainly data exfiltration. Most recent scan 2026-02-07.
This HeyGen Video Agent skill is GENERALLY SAFE with minor security considerations. The skill provides a straightforward wrapper around HeyGen's Video Agent API for generating AI avatar videos. The code uses standard Python libraries (urllib, json, pathlib) without dangerous operations like eval/exec. Network calls are limited to the legitimate HeyGen API endpoints (api.heygen.com). The skill requires an API key via environment variable, which is standard practice but requires user awareness. File operations are limited to creating output directories and downloading videos to user-specified locations. The referenced prompt-optimizer.md file is an internal resource bundled with the skill (normal behavior) and contains legitimate guidance for writing effective video generation prompts. No prompt injection, data exfiltration, command injection, or social engineering threats were detected. The main considerations are proper API key management and the standard risks of executing Python scripts that make network calls and write files.
Scanner's overall assessment (LLM analyzer)
Second scanner: NVIDIA SkillSpector
The two scanners disagree: Cisco found nothing and SkillSpector says do not install. One of them has a blind spot here; read the skill before trusting either. How the two scanners compare across the whole corpus: Two scanners, one corpus.
What this skill does
Generate AI avatar videos with HeyGen's Video Agent API. One-shot prompt to video generation.
- Package
- 4 files, 30 KB
- SKILL.md
- 290 words, 10 sections
- Code
- Python (2)
SKILL.md covers: HeyGen Video Agent · Reference Files · Setup · Generate a Video · Quick submit (returns video_id immediately) · Wait for completion and get video URL · Wait and download the video · Check Video Status · API Reference · Notes
Findings
MEDIUM API Key Required in Environment Variable Data Exfiltrationllm
The skill requires HEYGEN_API_KEY to be set as an environment variable. While this is a standard practice for API authentication, users should be aware that the scripts will have access to this credential. The skill does not implement any additional security measures like key validation, rotation reminders, or scope limitations.
SKILL.md
export HEYGEN_API_KEY="your-api-key"What is in the package
| File | Kind | Size |
|---|---|---|
SKILL.md | doc | 2.6 KB |
references/prompt-optimizer.md | doc | 19 KB |
scripts/generate.py | script | 4.6 KB |
scripts/status.py | script | 4.3 KB |
External hosts referenced
api.heygen.com app.heygen.com docs.heygen.com
Hostnames that appear in the skill's files. Referencing a host is not a finding; compare against the findings above and the skill's stated purpose.
Environment variables referenced
HEYGEN_API_KEY
Package contents recorded 2026-10-22. The scan above ran against the version current at scan time; the two can differ if the skill has since been updated.
What we checked
- Static analyzer. Pattern rules over SKILL.md and bundled files: known-bad commands, exfiltration URLs, secrets, obfuscation.
- Llm analyzer. A language model reads the skill and reports intent-level issues the rules cannot express.
- Meta analyzer. A second model pass that validates, deduplicates and re-scores the other analyzers' findings.
What we did not check
- Runtime behaviour. The skill was not executed; findings come from reading its files.
- Anything fetched at install or run time, including remote scripts, package dependencies, and URLs the skill tells the agent to open.
- Content the analyzers cannot read: images, binaries, and some non-English text.
- Known scanner bypass classes. See the methodology and limitations page.
Sources
- Repository: https://clawhub.ai/skills/video-agent
- Raw scan result: video-agent-scan.json
- Badge for your README:
- Request a re-scan