Skip to content

Latest commit

 

History

History
134 lines (97 loc) · 3.86 KB

File metadata and controls

134 lines (97 loc) · 3.86 KB

StackForge Release Readiness

This guide defines what must be true before StackForge should be presented as a reviewable V1 scaffold CLI.

CLI-first architecture

StackForge is intentionally CLI-first.

  • stackforge is the deterministic engine that lists templates, previews scaffold output, and writes repositories locally.
  • Humans and agents interact with the same command surface.
  • CrewCmd is a later wrapper, not the source of truth for generation logic.
  • OpenClaw agents should call the CLI only on machines where StackForge is already installed and verified.

That boundary keeps the core product testable without depending on orchestration state.

V1 commands

Release readiness for V1 is anchored to a small command surface:

stackforge --version
stackforge templates
stackforge init oss-cli demo --dry-run
stackforge init oss-cli demo

Expected behavior:

  • --version returns the installed CLI version.
  • templates lists the currently supported templates.
  • init ... --dry-run prints a deterministic plan without writing files.
  • init ... writes the scaffold locally and avoids hidden network steps.

launch can remain documented as a follow-on workflow, but V1 readiness should not depend on CrewCmd-driven orchestration being complete.

Command examples

Inspect available templates

pnpm dev templates

Preview a scaffold without writing files

pnpm dev init oss-cli my-tool --dry-run

Generate a local scaffold

pnpm dev init oss-cli my-tool

Release verification commands

pnpm check
pnpm build
pnpm check:templates
pnpm smoke:init

Safety model

StackForge should be safe by default.

  • No hidden network calls.
  • No implicit GitHub repository creation.
  • No default LLM calls.
  • No autonomous merging or publishing.
  • File generation must stay deterministic and reviewable.
  • Destructive overwrite behavior must require explicit operator intent.

The release bar is not just "it scaffolds"; it must scaffold in a way that is predictable enough for humans and agents to trust.

OpenClaw agent workflow

OpenClaw is an execution environment around StackForge, not a replacement for it.

  1. Verify the node or host has the StackForge CLI installed.
  2. Run stackforge templates or a dry run first to confirm the requested path.
  3. Use stackforge init ... --dry-run in planning or review-heavy flows.
  4. Run stackforge init ... only after the target repo name and template are confirmed.
  5. Hand the generated repository back to the normal branch, commit, PR, and review workflow.

This keeps agent usage grounded in the same deterministic CLI behavior a human maintainer would use locally.

CrewCmd-later boundary

CrewCmd support is valuable, but it is explicitly a later wrapper.

For V1, release-ready means:

  • the CLI works on its own
  • commands are documented
  • safety constraints are documented
  • agent usage expectations are documented
  • verification scripts pass

It does not require:

  • CrewCmd packaging to be complete
  • orchestration-specific automation to be the only entry point
  • StackForge to replace OpenClaw, CrewCmd, or GitHub review flows

V1 release checklist

Use this checklist with the general release process docs before opening a V1 release PR.

  • README explains the CLI-first architecture and CrewCmd-later boundary.
  • README or docs include working command examples for templates and init.
  • V1 command surface is limited to documented, supported behavior.
  • Safety expectations are documented: local-first, no hidden network calls, no implicit GitHub creation, no autonomous merge.
  • OpenClaw agent workflow is documented for dry runs and local execution.
  • pnpm check passes.
  • pnpm build passes.
  • pnpm check:templates passes.
  • pnpm smoke:init passes.
  • Any deferred CrewCmd work is clearly marked as post-V1.