The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes atlan CLI commands, such as atlan app init -o <app_path>, using variables resolved directly from user requests. In the absence of explicit instructions to sanitize or validate these inputs, this represents a potential vector for command injection if a user provides a maliciously crafted application name or path.
[EXTERNAL_DOWNLOADS]: The skill includes logic to install the atlan CLI if it is missing, specifying the use of Homebrew or pre-built binaries. These resources are associated with the vendor 'atlanhq' and represent standard environment preparation for the skill's intended functionality.
[PROMPT_INJECTION]: The skill is subject to an indirect prompt injection surface through its 'progressive discovery' workflow.
Ingestion points: The agent reads existing repository files, including main.py, workflow.py, and pyproject.toml, to inform its implementation strategy.
Boundary markers: The skill does not provide the agent with specific delimiters or instructions to ignore or isolate potential instructions embedded within the files it processes.
Capability inventory: The skill possesses the capability to execute CLI commands, write files to the system, and trigger external installation workflows.
Sanitization: No logic is provided to sanitize or validate the content extracted from repository files before it is used to influence the agent's implementation logic.

atlan-app-scaffold-standard