The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes the fentaris CLI for project initialization (fentaris init), configuration validation (fentaris check), and runtime diagnostics (fentaris doctor). It also facilitates the execution of package manager scripts such as build and typecheck.
[EXTERNAL_DOWNLOADS]: The skill requires the installation of the @fentaris/core library and related dependencies from the NPM registry as part of the project setup workflow.
[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection (Category 8) because it ingests user-provided data to configure external MCP servers and user identities.
Ingestion points: User responses to discovery questions (e.g., server names, transport commands, auth headers) documented in references/discovery.md.
Boundary markers: No explicit delimiters are used when interpolating user-provided strings into the generated TypeScript entrypoints or shell commands.
Capability inventory: The skill performs file system writes, shell command execution via the fentaris CLI, and facilitates network operations through the generated proxy configuration.
Sanitization: While the instructions recommend using the --non-interactive flag for automation safety, there is no explicit instruction to sanitize or validate the content of user-provided configuration values before they are used in code generation or CLI arguments.

fentaris-project-setup