The Agent Skills Directory

[PROMPT_INJECTION]: Potential indirect prompt injection vulnerability surface where untrusted data could influence command execution.
Ingestion points: Untrusted data enters the context through the $ARGUMENTS variable (parameters: <name>, --to <target>, --category <category>) and via user responses gathered during the interactive design phase described in SKILL.md.
Boundary markers: The instructions do not utilize delimiters or specific boundary markers to separate user-provided data from the shell command structure.
Capability inventory: The skill instructions involve executing subprocess commands including ls and the aiwg command-line utility.
Sanitization: There are no steps or instructions provided to validate, filter, or escape user-supplied strings before they are interpolated into shell-executed commands.
[COMMAND_EXECUTION]: The skill instructs the agent to perform directory listings and execute the aiwg scaffolding tool using arguments derived from user input. This pattern presents a risk of command injection if a user provides input containing shell metacharacters.

add-template