internal-linking
Pass
Audited by Gen Agent Trust Hub on Nov 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill does not contain any malicious instructions, credential harvesting, or unauthorized network operations. External links provided, such as those to LinkDR, are for documentation or related tools and originate from sources consistent with the author's identity.
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a provided utility script (
scripts/internal-link-graph.mjs) and shell one-liners usingnode -eandrg(ripgrep). These operations are restricted to the local repository for the purpose of identifying page routes and link placement. The code for these operations is transparently included in the skill body. - [INDIRECT_PROMPT_INJECTION]: The skill analyzes and modifies repository files (MDX, TSX, Astro) which are technically untrusted external data. While it lacks explicit boundary markers or sanitization for these inputs, the risk is inherent to any tool designed for code/content editing, and the operations performed (regex-based link matching) are low-risk.
- [OBFUSCATION]: A hex escape (
\x27) is used in a shell one-liner to represent a single quote. This is a common practice for escaping quotes in command-line contexts and does not represent an attempt to hide malicious behavior.
Audit Metadata