Security Audit
RightNow-AI/openfang:crates/openfang-hands/bundled/browser
github.com/RightNow-AI/openfangTrust Assessment
RightNow-AI/openfang:crates/openfang-hands/bundled/browser received a trust score of 85/100, placing it in the Mostly Trusted category. This skill has passed most security checks with only minor considerations noted.
SkillShield's automated analysis identified 2 findings: 0 critical, 0 high, 0 medium, and 1 low severity. Key findings include Potential for Credential Storage in Memory, Broad Web Access Capability.
The analysis covered 4 layers: Manifest Analysis, Static Code Analysis, Dependency Graph, LLM Behavioral Safety. All layers scored 70 or above, reflecting consistent security practices.
Last analyzed on February 27, 2026 (commit 7bd01856). SkillShield performs automated 4-layer security analysis on AI skills and MCP servers.
Layer Breakdown
Behavioral Risk Signals
Security Findings2
| Severity | Finding | Layer | Location | |
|---|---|---|---|---|
| LOW | Potential for Credential Storage in Memory The skill's 'Security Checklist' explicitly instructs the LLM to 'Never store passwords in memory_store'. This instruction highlights the potential for the `memory_store` tool (if available to the LLM) to be used for storing sensitive credentials, which could lead to credential harvesting if the LLM were to deviate from this instruction. While the skill provides a direct mitigation, the underlying capability for sensitive data storage exists. Ensure the LLM strictly adheres to the instruction not to store sensitive credentials in `memory_store`. Implement robust safeguards within the LLM's architecture to prevent it from storing or exfiltrating sensitive data, especially credentials, regardless of skill instructions. | LLM | SKILL.md:89 | |
| INFO | Broad Web Access Capability The skill utilizes browser automation tools (`browser_navigate`, `browser_read_page`, `browser_type`, `browser_click`) which inherently grant broad access to arbitrary web content. While the skill includes security advice like 'Verify domain before entering credentials' and 'Check for HTTPS', the underlying capability allows interaction with any website, including potentially malicious ones, if the LLM is instructed to do so or makes an error in navigation. This is an inherent risk of browser automation skills. Users should be aware of the broad web access capabilities and exercise caution when instructing the LLM to interact with unfamiliar websites. The LLM should be designed with strong guardrails to prevent navigation to known malicious sites or interaction with suspicious elements, and to strictly follow the skill's security checklist. | LLM | SKILL.md:70 |
Scan History
Embed Code
[](https://skillshield.io/report/a8f645e362417951)
Powered by SkillShield