The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted user-provided architectural designs and requirements in its 'Design Review' and 'Design Application' modes.
Ingestion points: User-submitted architectural descriptions and design proposals are ingested in SKILL.md for analysis.
Boundary markers: Absent. The skill does not define clear delimiters or specialized instructions to ignore embedded directives in user-provided content.
Capability inventory: The agent reads local documentation files and can execute the provided scripts/new_design.py tool. Malicious instructions in a design proposal could attempt to influence the agent's evaluation or tool usage.
Sanitization: Absent. There are no instructions for the agent to sanitize or validate external content before processing.
[COMMAND_EXECUTION]: The skill includes a Python utility, scripts/new_design.py, which is capable of writing content to the local filesystem.
Evidence: The script uses pathlib.Path.write_text() to save generated design documents to a user-specified path. If executed by an agent with tool privileges, a malicious user could potentially provide an output path that leads to arbitrary file writes or overwrites within the agent's working directory.

system-design-interview