github-explorer
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources such as GitHub Issues, Reddit, and various technical blogs. This represents an attack surface where an attacker could place malicious instructions in a GitHub issue or a forum post to influence the agent's output.
- Ingestion points: The skill uses
web_fetchandbrowsertools to retrieve content from GitHub repositories, issues, and third-party community platforms (Zhihu, V2EX, Reddit, etc.). - Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore potential commands found within the fetched content.
- Capability inventory: The skill's capabilities are limited to searching, fetching, and formatting data. It does not perform sensitive file operations, persistent configuration changes, or arbitrary shell command execution.
- Sanitization: There is no evidence of specific sanitization or filtering of the fetched content beyond the agent's natural summarization process.
Audit Metadata