The Agent Skills Directory

[COMMAND_EXECUTION]: The skill operates primarily through shell command execution using npx -y @insforge/cli. It also facilitates the installation and execution of external utilities like flyctl for compute management and the @posthog/wizard for analytics setup.
[EXTERNAL_DOWNLOADS]: Fetches deployment setup scripts from the vendor's repository and downloads the installation script for the well-known flyctl utility from Fly.io.
[REMOTE_CODE_EXECUTION]: Provides functionality to deploy user-authored edge functions to Deno Subhosting and containerized applications to Fly.io infrastructure.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources which creates an attack surface for indirect prompt injection.
Ingestion points: Reads database query results (db query), backend logs (logs), agent memory (memory recall), and allows the passing of issue descriptions to a backend diagnostic AI (diagnose --ai).
Boundary markers: The skill does not define explicit delimiters to separate untrusted data from instructions in the agent context.
Capability inventory: The tool can execute shell commands via npx, write migration files to the local file system, and perform network deployments to cloud providers.
Sanitization: The documentation notes that the CLI performs pattern-based redaction of sensitive patterns in logs and feedback, though it characterizes this as a safety net rather than a complete filter.
[DATA_EXFILTRATION]: Includes a feedback mechanism that transmits diagnostic data, including command history and error output, to the vendor's cloud infrastructure for support purposes.

insforge-cli