qmd
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEDYNAMIC_CONTEXT_INJECTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [DYNAMIC_CONTEXT_INJECTION]: The SKILL.md file contains a status check using the
!qmd statussyntax, which executes at load time. This is a benign implementation used to inform the agent if the necessary CLI tool is installed. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to search and retrieve data from local markdown notes, which may contain untrusted instructions.
- Ingestion points: Data is ingested through the search results and document retrieval functionalities (
structured_search,get,multi_get). - Boundary markers: The instructions do not specify any delimiters or safety prompts to isolate retrieved content from the system prompt.
- Capability inventory: The skill possesses capabilities for local file retrieval and search operations via the
qmdcommand. - Sanitization: No sanitization or filtering of retrieved markdown data is implemented.
- [EXTERNAL_DOWNLOADS]: The skill documentation refers to the
@tobilu/qmdpackage from the NPM registry as a required dependency. - [COMMAND_EXECUTION]: The skill explicitly allows the execution of
qmdCLI commands to perform index queries, health checks, and document retrieval.
Audit Metadata