1316e03b84
* chore(#4424): assert launcher snippet env surface is covered by scrub lists SNIPPET_SCRUB is hand-maintained for vars TEST_ENV_BASE's registry-derived list can't carry. Nothing asserted the union actually covers every ${VAR:-default} arm in _runtime-launcher.snippet.sh, so a new runtime-home arm with no scrub entry could drift silently — the #4205 shape, one door over. Adds (A2): extracts every ${[A-Z_]+:-} capture from the snippet and checks membership in TEST_ENV_BASE, SNIPPET_SCRUB, or the two vars the snippet/fixtures set themselves (RUNTIME_DIR, GSD_TOOLS). * fix(#4424): allow digits in the (A2) fallback-var regex CodeRabbit review on fork PR #37: [A-Z_]+ silently drops any \${VAR:-...} capture whose name contains a digit (e.g. CLAUDE2_CONFIG_DIR) instead of flagging it uncovered, defeating the guard's own purpose. Matches bash identifier syntax instead: leading letter/underscore, then alnum/underscore. * fix(#4424): rename SELF_ASSIGNED to reflect RUNTIME_DIR's real provenance Gemini adversarial review (agy) on fork PR #37: RUNTIME_DIR is an external input the snippet reads via \${RUNTIME_DIR:-...}, never assigns — every fixture sets it in-script before sourcing the snippet. Only GSD_TOOLS is truly snippet-self-assigned. SELF_ASSIGNED conflated the two; renamed to CALLER_OR_SELF_ASSIGNED. No behavior change. Reviewed and rejected: moving RUNTIME_DIR into SNIPPET_SCRUB (blanking it is indistinguishable from unset to the resolver's own \${RUNTIME_DIR:-...} fallback, re-opening the #4205 ambient-leak this suite guards against — see the existing comment at line ~1801); widening the regex to mixed-case, colon-less \${VAR-default}, or \${VAR:=default} forms (none exist in the snippet, and the issue's own spec scopes this to \${[A-Z_]+:- captures); stripping bash comments before matching (the snippet is one physical line with zero '#' characters, so no comment can exist in it). * fix(#4424): guard CALLER_OR_SELF_ASSIGNED against silent future additions trek-e review on PR #4503: a future ${VAR:-default} arm could be dropped into this set without confirming it is genuinely caller-supplied/ self-assigned rather than a real coverage gap. Adds a comment requiring justification for any addition, pointing to SNIPPET_SCRUB as the default when in doubt. No behavior change. * fix(#4424): guard (A2) against a vacuous pass on empty extraction agy adversarial review (gemini-3.8-flash-high, /gsd-review lane) on PR #4503: if the snippet becomes unreadable/truncated/renamed, matchAll yields zero matches, uncovered stays [], and assert.deepStrictEqual passes vacuously — same "guards the guard" gap the sibling (E)-adjacent tests already close with assert.ok(files.length > 0, ...). Asserts extracted.length >= 15 before filtering. Reviewer's second finding (regex misses colon-less ${VAR-default}) is not applied: no such form exists in the snippet today, and 688cc1c already recorded this exact widening as scope creep the issue's own spec (${[A-Z_]+:-) does not ask for. --------- Co-authored-by: Test <test@test.com> Co-authored-by: Tom Boucher <trekkie@nomorestars.com>
GSD Core
Git. Ship. Done.
English · Português · 简体中文 · 日本語 · 한국어
A light-weight meta-prompting, context engineering, and spec-driven development system for Claude Code, OpenCode, Antigravity CLI, Kimi CLI, Kilo, Codex, Copilot, Cursor, Windsurf, and more.
What is GSD Core
GSD Core is a context-engineering and spec-driven development framework that drives AI coding agents (Claude Code, Codex, Antigravity CLI, Kimi CLI, Copilot, Cursor, and more) through a disciplined phase loop. It solves context rot — the quality degradation that accumulates as an AI fills its context window — by running all heavy research, planning, and execution work in fresh-context subagents while keeping your main session lean.
How it works
Each milestone repeats the same five-step loop, one phase at a time:
- Discuss — capture implementation decisions before anything is planned
- Plan — research, decompose, and verify the plan fits a fresh context window
- Execute — run plans in parallel waves; each executor starts with a clean 200k-token context
- Verify — walk through what was built; diagnose and fix before declaring done
- Ship — create the PR, archive the phase, repeat for the next one
Quickstart
npx @opengsd/gsd-core@latest
The installer prompts for your runtime (Claude Code, OpenCode, Antigravity CLI, Kimi CLI, Kilo, Codex, Copilot, Cursor, Windsurf, and more) and whether to install globally or locally. The installer is required for cross-runtime compatibility — do not copy files from agents/ or commands/ directly.
On another runtime or without Node.js? See Install on your runtime.
Once installed, start a new project or onboard an existing repo:
/gsd-new-project # greenfield project
/gsd-onboard # existing codebase
New here? Follow Your first project for a guided walkthrough from install to first shipped phase, or Onboarding an existing codebase for brownfield setup.
Documentation
What's new in 1.7.0 → docs/whats-new-1.7.0.md
Tutorials — learning by doing:
How-to guides — task-focused recipes:
Reference — authoritative facts:
Explanation — concepts and design decisions:
Full index: docs/README.md. Other languages: 日本語 · 한국어 · Português · 简体中文.
Why it works
Most AI-coding setups fail at scale because context bloat silently degrades output quality, there is no shared memory between sessions, and nothing verifies that code actually works. GSD Core solves all three: heavy work runs in fresh subagents, structured artifacts like STATE.md and CONTEXT.md survive session boundaries, and the verify step walks through what was built and generates fix plans before a phase is declared done. See docs/explanation/context-engineering.md for the full reasoning.
Troubleshooting? See docs/how-to/recover-and-troubleshoot.md.
Community
| Project | Platform |
|---|---|
| gsd-opencode | Original OpenCode port |
| Discord | Community support |
Star History
License
MIT License. See LICENSE for details.
Claude Code is powerful. GSD Core makes it reliable.