markdown-formatter
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted markdown files, which presents a surface for indirect prompt injection (SKILL.md).
- Ingestion points: External markdown files read during the formatting workflow.
- Boundary markers: The instructions explicitly mandate wrapping content in
<untrusted-content>tags and ignoring any embedded directives such as "Ignore previous instructions". - Capability inventory: The agent has access to file system tools (Read, Edit, Grep, Glob) and a custom bash script.
- Sanitization: Input paths are validated in the validate-markdown.sh script.
- [COMMAND_EXECUTION]: The skill executes a local validation script using the Bash tool (SKILL.md, scripts/validate-markdown.sh).
- The execution pattern uses the -- separator to prevent filename-based argument injection.
- The script implements security checks including path traversal detection (..), file extension enforcement (.md), and a character whitelist for input filenames.
Audit Metadata