How to Understand ClawHub Security Audits Before Installing Skills or Plugins
This page explains how to read ClawHub security audits to assess whether a skill or plugin is safe to install. It covers audit status, risk levels, and key trust indicators for users evaluating releases.
Read this when
- Understanding ClawHub security audit results
- Deciding whether to install a skill or plugin
- Explaining ClawHub audit status, risk level, or findings
Security Audits
ClawHub security audits give you the information needed to assess whether a skill or plugin is safe to install. Each audit reveals what the release actually does, which permissions it requests, and whether any red flags exist before it can interact with files, accounts, passwords, source code, or third-party services.
Audits provide strong safety indicators, but they do not guarantee a release is completely safe. Always apply your own judgment before authorizing sensitive access.
Related topics: Security, Acceptable usage, and Moderation and Account Safety.
What to check before installing
Prior to installation, examine:
- the audit status summary
- the assigned risk level
- any findings that appear
- required credentials, permissions, or environment variables
- owner identity, source location, version number, changelog, download count, star rating, and other trust indicators
Only install content you fully understand and trust.
Audit status
The audit status tells you how to interpret the audit result:
| Status | Meaning |
|---|---|
Pass | No issue above low risk was detected. |
Review | Review the findings before installing. The release could still be legitimate. |
Warn | Proceed with extra caution. ClawHub identified a high-impact concern or warning. |
Malicious | Do not install. |
Pending | Audits are still in progress. |
Error | The audit could not finish. |
A Pass status is reassuring, but it does not replace your own evaluation. This is especially important for tools that can publish content, modify data, execute commands, read files, or reach production systems.
Risk level
Risk level indicates blast radius: how much authority the release appears to have when used as intended.
| Risk level | Meaning |
|---|---|
Low | Little sensitive authority or user impact was found. |
Medium | The release carries meaningful authority, such as account access or data modification. |
High | The release has high-impact authority, severe findings, or malicious indicators. |
Risk level and audit status answer different questions:
- Risk level asks: "How much power does this have?"
- Audit status asks: "What action should I take based on this result?"
For instance, a publishing skill might show Review with Medium risk. That does not mean it is malicious. It means the skill appears aligned with its purpose but can exercise meaningful account authority.
Findings
Findings explain why a particular audit result was produced. Each finding typically includes:
- what it means
- why it was flagged
- the relevant skill or plugin content
- a recommendation
Findings may be labeled Info, Low, Medium, High, or Critical. Higher severity findings contribute more heavily to both risk level and audit status.
Low-confidence findings are excluded from the public audit summary to keep the page focused on actionable evidence.
What ClawHub checks
ClawHub audits submitted release artifacts, including:
- skill instructions or plugin metadata
- declared environment variables and permissions
- installation instructions and package metadata
- included files and file manifests
- compatibility and capability metadata
The central question is coherence: do the name, summary, metadata, requested authority, and actual content match what users would reasonably expect?
Powerful behavior is not automatically problematic. Many useful tools require credentials, local commands, provider APIs, or package installations. The audit checks whether that power is expected, disclosed, and proportional.
Artifact pages link to the complete audit at:
/<owner>/skills/<slug>/security-audit
The audit page combines:
- SkillSpector
- VirusTotal
- Risk analysis
VirusTotal
ClawHub incorporates VirusTotal as malware telemetry within the audit pipeline. VirusTotal is a trusted industry standard for file reputation and malware scanning, and our partnership allows ClawHub to add broader security intelligence to skill and plugin review.
VirusTotal is particularly useful for known malicious artifacts, engine hits, and reputation signals that supplement ClawHub's agent-aware review. When vendor engine counts are available, the audit summarizes them in plain language, for example:
62/62 vendors flagged this skill as clean.
or:
2/64 vendors flagged this skill as malicious, 1/64 flagged it as suspicious, and 61/64 flagged it as clean.
When ClawHub has no vendor-count telemetry to summarize, the audit states:
No VirusTotal findings
VirusTotal remains telemetry. It does not replace ClawHub's own artifact-aware risk analysis.
Risk analysis
Risk analysis is powered internally by ClawScan, ClawHub's own security audit system. It reviews each release as an agent-facing artifact: instructions, metadata, declared permissions, files, capability signals, static scan signals, SkillSpector findings, VirusTotal telemetry, and publisher-provided context. Static scan signals are internal context for this review; they are not a standalone public audit section or install-blocking verdict.
Risk analysis uses the OWASP Agentic Skills Top 10 as a lens for risks such as prompt injection, tool misuse, credential exposure, unsafe execution, memory or context poisoning, and excessive agency.
ClawScan does not treat a scary-looking capability as automatically malicious. It asks whether the capability is disclosed, purpose-aligned, and supported by the release's stated use case.