baoyu-format-markdown
Pass
Audited by Gen Agent Trust Hub on Mar 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes shell commands for file management and script execution. Evidence includes using the
bashtool inSKILL.mdfor backups viamvand runningnpx -y bun. Inscripts/autocorrect.ts,execSyncrunsnpx autocorrect-nodeusing the file path as an argument. - [EXTERNAL_DOWNLOADS]: The skill uses
npxto runbunandautocorrect-node, which may download these packages from the NPM registry if not present. These are well-known developer tools. - [PROMPT_INJECTION]: The skill processes untrusted file content, providing a surface for indirect prompt injection. Ingestion points: Reads files via the
readtool andreadFileSync. Boundary markers: Absent. Capability inventory: Includesbashfor backups and script execution, andwritefor saving files. Sanitization: No explicit sanitization of text content before language model analysis.
Audit Metadata