Trust Assessment
sglang received a trust score of 53/100, placing it in the Caution category. This skill has some security considerations that users should review before deployment.
SkillShield's automated analysis identified 5 findings: 0 critical, 2 high, 2 medium, and 1 low severity. Key findings include Network egress to untrusted endpoints, Covert behavior / concealment directives, Unpinned dependency in manifest.
The analysis covered 4 layers: Manifest Analysis, Static Code Analysis, Dependency Graph, LLM Behavioral Safety. The Static Code Analysis layer scored lowest at 63/100, indicating areas for improvement.
Last analyzed on February 12, 2026 (commit 458b1186). SkillShield performs automated 4-layer security analysis on AI skills and MCP servers.
Layer Breakdown
Behavioral Risk Signals
Security Findings5
| Severity | Finding | Layer | Location | |
|---|---|---|---|---|
| HIGH | Unpinned dependency in manifest The skill's manifest lists 'sglang', 'torch', and 'transformers' as dependencies without specifying exact versions. This can lead to non-deterministic builds and introduce vulnerabilities if a new version of a dependency contains a breaking change or a security flaw. An attacker could potentially publish a malicious package under one of these names, which would then be installed. Pin all dependencies to exact versions (e.g., 'sglang==1.0.0', 'torch==2.3.0'). Use a dependency management tool that generates lock files. | Static | Manifest | |
| HIGH | Unpinned dependency in installation instructions The `pip install "sglang[all]"` and `pip install sglang[all] flashinfer` commands do not specify exact versions for `sglang` or `flashinfer`. This can lead to non-deterministic builds and introduce vulnerabilities if a new version of a dependency contains a breaking change or a security flaw. An attacker could potentially publish a malicious package under one of these names, which would then be installed. Specify exact versions for all dependencies in installation instructions (e.g., `pip install "sglang[all]==1.0.0"`). | Static | SKILL.md:39 | |
| MEDIUM | Network egress to untrusted endpoints HTTP request to raw IP address Review all outbound network calls. Remove connections to webhook collectors, paste sites, and raw IP addresses. Legitimate API calls should use well-known service domains. | Manifest | cli-tool/components/mcps/devtools/figma-dev-mode.json:4 | |
| MEDIUM | Installation from unpinned Git source The installation instructions include `git clone https://github.com/sgl-project/sglang.git` followed by `pip install -e "python[all]"`. Cloning from a Git repository without specifying a particular commit hash or tag means that the installed code can change over time. If the `main` branch (or default branch) of the repository is compromised, a malicious version of the code could be installed. Installing in editable mode (`-e`) further increases the risk as changes in the source directory are immediately reflected. If installing from source is necessary, specify a fixed commit hash or tag (e.g., `git clone ... && cd sglang && git checkout <commit_hash>`). For production, prefer installing from a trusted package index with pinned versions. | Static | SKILL.md:45 | |
| LOW | Covert behavior / concealment directives Multiple zero-width characters (stealth text) Remove hidden instructions, zero-width characters, and bidirectional overrides. Skill instructions should be fully visible and transparent to users. | Manifest | cli-tool/components/mcps/devtools/jfrog.json:4 |
Scan History
Embed Code
[](https://skillshield.io/report/1fe00f40f25eb9dc)
Powered by SkillShield