Security audit
Security checks for vulnerabilities and agentic risk
This skill fetches GitHub repository files for analysis, and its network access is disclosed and aligned with that purpose.
Install this if you want agents to fetch and analyze GitHub repositories with curl. Be aware that repository names, branches, and paths you ask about will be requested from GitHub, and fetched repository text should still be treated as untrusted content rather than executed code.
The activation criteria are broad enough to trigger on many ordinary requests about GitHub URLs or repository understanding, which can cause the skill to run in contexts the user did not explicitly intend. Because the skill then directs the agent to make outbound network requests, overbroad triggering increases the chance of unnecessary external access and unintended data handling.
The skill explicitly instructs the agent to send data to external GitHub endpoints using curl. Even though GitHub is the intended destination, automatic outbound requests can disclose user-supplied repository targets, consume network privileges, and fetch untrusted remote content without an explicit consent or safety gate; the 'ALWAYS use curl' wording makes this more dangerous by removing discretion.
# 1. List root directory
curl -s "https://api.github.com/repos/OWNER/REPO/contents/"
# 2. Get README
curl -s "https://raw.githubusercontent.com/OWNER/REPO/main/README.md"
This instruction expands the external-fetch workflow to additional repository paths based on prior findings, which can lead to iterative remote retrieval of arbitrary repository content. In context, the skill is designed for repository analysis, so some network access is expected, but the unqualified fetch-first behavior still increases exposure to unnecessary outbound transmission and retrieval of adversarial content.
curl -s "https://raw.githubusercontent.com/OWNER/REPO/main/README.md"
curl -s "https://api.github.com/repos/OWNER/REPO/contents/src"
## Get specific file
The description includes Chinese trigger examples alongside English instructions, but it does not state whether responses should match the user's language or offer any language preference choice. This can create an implicit language behavior without explicit user opt-in or documented locale handling.
No suspicious patterns detected.