The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/generate_brief.sh contains a command injection vulnerability. The functions prompt_input and prompt_multiline utilize eval to assign user-provided input to shell variables without sanitization (e.g., eval "$var_name=\"$input\""). This allows execution of arbitrary commands if the input contains shell substitutions or metacharacters.
[DYNAMIC_EXECUTION]: The skill performs unsafe dynamic execution by using eval on string content built from external user input in scripts/generate_brief.sh. Malicious data provided to the interactive prompts could be executed as shell commands.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates indirect prompt injection by ingesting untrusted data through an interactive script and persisting it into a markdown template. 1. Ingestion points: User input captured via scripts/generate_brief.sh prompts for audience, problem statement, and requirements. 2. Boundary markers: Absent. The resulting brief files do not include delimiters or instructions to treat the content as untrusted data. 3. Capability inventory: The skill can perform filesystem writes to create the brief documents. 4. Sanitization: Absent. Input is directly interpolated into the markdown output via heredoc without escaping or validation.

content-brief-generator