The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill documentation includes an installation command that fetches a script from the official .NET Aspire domain and executes it via the shell. This is a standard pattern for CLI distribution and originates from a trusted service.\n
Evidence: curl -sSL https://aspire.dev/install.sh | bash in SKILL.md.\n- [COMMAND_EXECUTION]: The skill uses the aspire CLI to perform administrative and orchestration tasks, such as starting and stopping background processes, managing resource state, and diagnosing environment issues.\n
Evidence: Commands like aspire start, aspire stop, and aspire resource rebuild are used throughout the instruction set.\n- [PROMPT_INJECTION]: The agent is instructed to parse output from CLI commands that aggregate metadata from the user's project files. This creates a surface for indirect prompt injection if project definitions or resource names contain adversarial instructions intended to influence the agent's behavior.\n
Ingestion points: aspire ps --format Json, aspire describe --format Json (reads resource names and states from project files).\n
Boundary markers: Instructions suggest using JSON format and stripping non-JSON lines to handle mixed output, but do not explicitly enforce safety delimiters or warnings for the resource data content itself.\n
Capability inventory: The skill utilizes subprocess calls for process management (aspire start/stop) and file system operations through the CLI.\n
Sanitization: Not explicitly implemented; the agent relies on structured JSON parsing of the command output.

aspire-orchestration