The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it is designed to ingest and process untrusted data from external sources.
Ingestion points: The workflow instructions specify reading from ai/memory-bank/site-setup.md and ai/memory-bank/tasks/*-tasklist.md using cat and grep.
Boundary markers: No specific delimiters or instructions to ignore embedded commands within the ingested files are present.
Capability inventory: The agent has the capability to generate code (CSS, JS, HTML) and design specifications based on these inputs.
Sanitization: There is no evidence of sanitization or filtering of the content read from the memory bank files before it is processed by the LLM.
[COMMAND_EXECUTION]: The skill utilizes local shell commands as part of its defined workflow.
Evidence: The workflow section in SKILL.md includes commands such as cat ai/memory-bank/site-setup.md and grep -i "target\|audience\|goal\|objective" ai/memory-bank/site-setup.md to gather context. These are standard utility commands for reading project documentation and do not involve high-risk operations like downloading or executing remote scripts.

design-ux-architect