sync-docs
Pass
Audited by Gen Agent Trust Hub on Jul 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill defines a shell-based discovery step using
find <target> -name "*.md" -type f. The use of an unvalidated<target>parameter represents a command injection surface if the input is not sanitized by the executing agent. - [PROMPT_INJECTION]: The skill processes untrusted content from documentation and code files which are then used to instruct sub-agents. This creates a surface for indirect prompt injection.
- Ingestion points: Markdown documentation files and implementation source code read during Phase 1 and Phase 2.
- Boundary markers: Sub-agent tasks lack explicit delimiters or safety instructions to ignore commands embedded in the project files.
- Capability inventory: The skill has the ability to run shell commands for file lookup and orchestrate sub-agents to modify file content.
- Sanitization: No filtering or sanitization of the input file content is implemented.
Audit Metadata