The Agent Skills Directory

[PROMPT_INJECTION]: The skill directly interpolates user input from the $ARGUMENTS variable into the system prompt without utilizing boundary markers (such as XML tags or triple quotes) to separate instructions from data. This creates a potential surface for indirect prompt injection where adversarial input could attempt to influence the agent's behavior. * Ingestion points: User-provided input via $ARGUMENTS in SKILL.md. * Boundary markers: Absent; input is directly interpolated into the system instructions. * Capability inventory: Text generation only; no specific tool calls or script executions are defined in the provided file. * Sanitization: Absent; no explicit instructions to escape or validate the input data.
[NO_CODE]: The skill consists entirely of instructional markdown and metadata, and does not include any executable scripts or binary files, which inherently limits the potential for malicious code execution outside the prompt context.

business-model