dockerfile-gen
Verdict safe, but 2 findings (2 medium), mainly social engineering, data exfiltration. Most recent scan 2026-02-06.
This skill has MEDIUM severity security concerns primarily related to deceptive presentation and lack of transparency. The core issue is that this is not actually an agent skill with executable code, but rather documentation for an external npm package (ai-dockerfile) that requires Node.js and OpenAI API credentials. Users may be misled into: (1) believing this is a native agent capability when it's actually an external tool, (2) exposing OpenAI API credentials to a third-party npm package without understanding the implications, and (3) unknowingly transmitting project source code to OpenAI APIs. The skill lacks transparency about data flows, credential usage, and external dependencies. While no active malicious behavior is detected, the deceptive presentation and undisclosed external dependencies create security risks through user confusion and uninformed credential exposure.
Scanner's overall assessment (LLM analyzer)
Second scanner: NVIDIA SkillSpector
Independent static run, no language model, over the same files as the Cisco scan above. How the two scanners compare across the whole corpus: Two scanners, one corpus.
What this skill does
Generate optimized Dockerfiles for any project
- Package
- 1 file, 1.2 KB
- SKILL.md
- 145 words, 12+ sections
- Code
- None. Instructions only.
SKILL.md covers: Dockerfile Generator · Quick Start · What It Does · Usage Examples · Generate for current project · Specify base image · Production optimized · With compose file · Output Includes · Features · Requirements · License
Findings
MEDIUM Misleading Skill Description - External Tool Masquerading as Agent Skill Social Engineeringllm
The skill presents itself as an agent skill that generates Dockerfiles, but the instructions exclusively reference 'npx ai-dockerfile', which is an external npm package/CLI tool, not agent-executable code. The skill contains no Python/Bash scripts for the agent to execute. This is deceptive because users expect agent skills to provide capabilities the agent can directly execute, not just documentation for external tools. The skill effectively tricks users into installing and running an external package under the guise of being a native agent capability.
SKILL.md
Instructions contain only 'npx ai-dockerfile' commands with no agent-executable scripts. Examples: 'npx ai-dockerfile', 'npx ai-dockerfile --base node:20-alpine', 'npx ai-dockerfile --production'. No Python/Bash files present in skill package.MEDIUM Undisclosed External API Dependency and Credential Requirement Data Exfiltrationllm
The skill requires OPENAI_API_KEY environment variable but this critical requirement is buried in the documentation and not declared in the YAML manifest. Users may unknowingly expose their OpenAI API credentials to an external npm package without understanding the data flow or security implications. The skill does not disclose what data is sent to OpenAI APIs, whether project source code is transmitted, or how credentials are handled by the external tool.
SKILL.md:45
'Requirements: Node.js 18+. OPENAI_API_KEY required.' - This critical credential requirement is mentioned only in prose, not in manifest metadata. No disclosure of what data is sent to OpenAI or how the API key is used.What is in the package
| File | Kind | Size |
|---|---|---|
SKILL.md | doc | 1.2 KB |
External hosts referenced
github.com x.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.
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.
- Behavioral analyzer. Checks what the skill asks the agent to do against what its description claims it does.
- Llm analyzer. A language model reads the skill and reports intent-level issues the rules cannot express.
- Trigger analyzer. Looks for instructions that fire on conditions the user did not ask for.
- 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/ai-dockerfile
- Raw scan result: dockerfile-generator-scan.json
- Badge for your README:
- Request a re-scan