hugging-face-cli
Verdict safe, but 5 findings (5 low), mainly data exfiltration, policy violation, social engineering. Most recent scan 2026-02-07.
The hugging-face-cli skill is a documentation-focused skill that provides command reference and examples for the Hugging Face CLI tool. The skill does not contain executable scripts (Python/Bash) and serves primarily as a reference guide. Security analysis reveals only minor informational issues: missing optional metadata fields (allowed-tools, license, compatibility), missing referenced files, and opportunities to enhance security guidance around token handling. No critical or high-severity threats were identified. The skill does not attempt prompt injection, data exfiltration, command injection, or other malicious behaviors. The internal reference files (references/commands.md, references/examples.md) contain legitimate documentation and examples without malicious content. Overall, this skill presents minimal security risk and functions as intended - a reference guide for Hugging Face CLI operations.
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
Execute Hugging Face Hub operations using the `hf` CLI. Use when the user needs to download models/datasets/spaces, upload files to Hub repositories, create repos, manage local cache, or run compute jobs on HF infrastructure. Covers authentication, file transfers, repository creation, cache operations, and cloud compute.
- Package
- 3 files, 42 KB
- SKILL.md
- 950 words, 12+ sections
- Code
- None. Instructions only.
SKILL.md covers: Hugging Face CLI · Quick Command Reference · Core Commands · Authentication · Download · Upload · Repository Management · Delete Files from Repo · Cache Management · Browse Hub · Models · Datasets
Findings
LOW Skill does not specify a license Policy Violationstatic
Skill manifest does not include a 'license' field. Specifying a license helps users understand usage terms.
SKILL.md
LOW Missing allowed-tools Declaration Data Exfiltrationllm
The skill does not declare 'allowed-tools' in its YAML manifest. While this field is optional per the agent skills specification, declaring it would improve transparency about which agent capabilities (Bash, Python, Read, Write, etc.) this skill requires. Since the skill executes Bash commands via the 'hf' CLI, it should ideally declare 'allowed-tools: [Bash]' for clarity.
SKILL.md
allowed-tools: Not specifiedLOW Missing License and Compatibility Metadata Social Engineeringllm
The skill does not specify 'license' or 'compatibility' fields in its YAML manifest. While these are optional fields, including them would improve transparency about licensing terms and which environments (Claude.ai, Claude Code, API) the skill is designed for.
SKILL.md
license: Not specified
compatibility: Not specifiedLOW Authentication Token Handling in Examples Data Exfiltrationllm
The skill documentation includes examples of passing authentication tokens via command-line arguments ('hf auth login --token $HF_TOKEN'). While using environment variables is a reasonable practice, the skill should emphasize secure token handling and warn against hardcoding tokens in scripts or command history.
hf auth login --token $HF_TOKEN # Non-interactive with tokenLOW Missing Referenced Files Data Exfiltrationllm
The skill references several files (cache.py, Repo.py, templates/examples.md, templates/commands.md, assets/examples.md, assets/commands.md) that were not found in the skill package. While two reference files (references/commands.md, references/examples.md) exist, the missing files suggest incomplete packaging or documentation references to non-existent resources.
Referenced files not found: cache.py, Repo.py, templates/examples.md, templates/commands.md, assets/examples.md, assets/commands.mdWhat is in the package
| File | Kind | Size |
|---|---|---|
SKILL.md | doc | 7.3 KB |
references/commands.md | doc | 27 KB |
references/examples.md | doc | 8.3 KB |
External hosts referenced
huggingface.co
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
HF_TOKEN MODEL_NAME ORG HF_HUB_DOWNLOAD_TIMEOUT HF_HUB_CACHE HF_HUB_DISABLE_PROGRESS_BARS HF_HUB_OFFLINE MODELS MODEL_PATH UPLOAD_URL WANDB_KEY
Package contents recorded 2026-10-22 at commit 2afeee6dc727. 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://github.com/huggingface/skills
- Raw scan result: skills-hugging-face-cli-scan.json
- Badge for your README:
- Request a re-scan