The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the global installation of a third-party NPM package '@tobilu/qmd' which is not from a verified or well-known technology vendor.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests and processes untrusted data from a document index. 1. Ingestion points: Results from 'qmd query' (SKILL.md). 2. Boundary markers: Absent; there are no instructions to use delimiters or ignore embedded instructions when presenting search results. 3. Capability inventory: Execution of 'qmd' commands and output presentation. 4. Sanitization: Absent; the content is presented directly from the tool's JSON output.
[COMMAND_EXECUTION]: The skill executes the 'qmd' CLI tool with user-supplied queries using string interpolation. This poses a command injection risk if the calling environment does not properly sanitize the query string before it is executed in the shell.

rag-search