readable-mdn
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches documentation content from the official
mdnorganization repositories on GitHub (raw.githubusercontent.com/mdn/contentandraw.githubusercontent.com/mdn/translated-content). These are well-known, trusted sources for web development documentation. - [COMMAND_EXECUTION]: The skill uses the
curlutility to download Markdown files to a temporary location (/tmp/mdn.md). The execution is limited to retrieving content from defined Mozilla documentation paths. - [INDIRECT_PROMPT_INJECTION]: The skill processes external documentation which could theoretically contain malicious instructions.
- Ingestion points: Remote content is downloaded via
curlfromraw.githubusercontent.com/mdnand read into the agent's context from/tmp/mdn.mdinSKILL.md. - Boundary markers: None explicitly defined in the instructions.
- Capability inventory: Uses
curlfor network access and file system operations (writing and reading in/tmp) as defined inSKILL.md. - Sanitization: No specific sanitization or filtering of the retrieved Markdown is performed, relying on the reputation of the official MDN repositories.
Audit Metadata