The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell scripts in SKILL.md to manage telemetry and workspace setup, executing commands like cat, grep, sed, mkdir, and curl to interact with the local environment.- [DATA_EXFILTRATION]: A telemetry system transmits usage metrics and platform information to an external endpoint (Convex) via curl. This behavior reads from ~/.superstack/config.json and is gated by a user consent prompt.- [REMOTE_CODE_EXECUTION]: The skill performs remote code operations, including cloning GitHub repositories and installing packages via npx skills add and npx create-solana-dapp. It references various repositories and templates from the Solana ecosystem.- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface with the following characteristics:
Ingestion points: Project requirements are read from .superstack/idea-context.md in SKILL.md.
Boundary markers: Absent; the skill does not use delimiters or instructions to ignore embedded commands when reading context data.
Capability inventory: The skill possesses the ability to execute shell commands, clone repositories, and install remote packages (SKILL.md).
Sanitization: Absent; the content from the context file is processed to inform project structure and dependency selection without explicit validation.

scaffold-project