The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute various subcommands of the stackbone CLI tool to manage project lifecycles. This includes initializing workspaces (stackbone init), adding components (stackbone add), running local development environments (stackbone dev), and publishing artifacts (stackbone publish). These operations are consistent with the tool's intended use for software development.
[EXTERNAL_DOWNLOADS]: The skill mentions downloading and installing software packages using npm, pnpm, and npx. Referenced packages include @stackbone/cli, @stackbone/sdk, and several @langchain libraries. These resources are from established vendors or well-known open-source projects and are retrieved from standard package registries.
[INDIRECT_PROMPT_INJECTION]: The skill's architecture involves ingesting data from external files and user-provided inputs, creating a potential surface for indirect prompt injection if untrusted content is processed.
Ingestion points: Data is brought into the environment via commands like stackbone rag ingest <path>, stackbone config set --file <path>, and stackbone prompts create --file <path> (referenced in references/rag.md, references/config.md, and references/prompts.md).
Boundary markers: The provided instructions do not specify the use of delimiters or 'ignore' instructions when handling the content of these ingested files.
Capability inventory: The skill enables broad capabilities including file system access, network communication with a control plane, and the execution of shell commands.
Sanitization: The documentation does not describe explicit sanitization or validation steps for the content of processed documents prior to ingestion into the agent's context.

stackbone-cli