The Agent Skills Directory

[PROMPT_INJECTION]: The skill utilizes the $ARGUMENTS variable to define the scope of the feature implementation. It does not employ boundary markers or explicit instructions to treat the user-supplied content strictly as data, creating a surface for indirect prompt injection.
Ingestion points: The feature description is ingested via $ARGUMENTS in the primary task definition within SKILL.md.
Boundary markers: No delimiters (such as XML tags or triple backticks) or 'ignore embedded instructions' warnings are present around the user input.
Capability inventory: The skill allows the agent to modify sensitive files across the entire stack (Database, API, Frontend) and execute shell commands.
Sanitization: There is no evidence of validation or sanitization of the provided arguments before they are used to guide the agent's multi-step implementation process.
[COMMAND_EXECUTION]: Step 6 of the skill explicitly directs the agent to execute shell commands (npm run lint-dev and npm run test-api). While these are common development tasks, they serve as a potential execution vector for malicious code that could be introduced if the agent is successfully manipulated via the indirect prompt injection surface mentioned above.

add-feature