30c40e0fe2
* test(#4302): add failing regression test for deferral-window paragraph leak RED: hasNearbyDeferralMarker's 500-char trailing window has no structural boundary, so a compact citation-free restatement immediately followed by an unrelated section that cites tdd.md borrows that neighbor's citation and evades restatesCycleStructurally(). Reproduces the exact shape confirmed via mutation testing against the real agents/gsd-executor.md. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(#4302): bound the deferral window at the next heading, not a flat char count hasNearbyDeferralMarker's 500-char trailing window had no structural boundary, so a compact citation-free restatement immediately followed by an unrelated section that cites tdd.md borrowed that neighbor's citation and evaded restatesCycleStructurally() (confirmed on the real gsd-executor.md, whose next section after the cycle pointer has its own independent citation 82 chars past the REFACTOR anchor). First attempt bounded at the next blank line instead, and was rejected after it broke a real case caught by direct execution before committing: execute-plan.md's FIRST RED/GREEN/REFACTOR occurrence is a numbered list's own intro sentence, separated by a genuine blank line from the list item that actually carries the citation — a blank line is not reliably "still the same statement" once list structure is involved. Bounds at the next markdown heading (`\n#`) instead, capped at 500 chars as before when no heading appears. Verified against both real files, the new regression fixture, and every pre-existing #4268 fixture/property test. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(#4302): close a narrower same-section evasion found by orthogonal review The heading-only fix left a narrower gap: a citation-free restatement followed by a blank line and an unrelated PROSE paragraph (no heading) that cites tdd.md for an unrelated reason still evaded detection — confirmed by the reviewer via execution. Generalizes the boundary rule: any blank line whose following line is NOT a markdown list continuation is now also a boundary (in addition to the existing heading boundary), reconciling with the earlier-rejected flat blank-line bound by adding the list-continuation exception that execute-plan.md's real shape (a numbered list's intro sentence, then a blank line, then the list item carrying the citation) needs. Verified against both real files, all three restatement fixtures (heading-bounded, prose-bounded, list-continuation negative-space), every pre-existing #4268 fixture, and the fast-check property test (200 runs). Also restores an honest, updated limitation disclosure describing the one narrower residual case not resolved by this design (a citation reachable via exactly one list-item hop from an unrelated list) — matching this repo's disclosed-not-silently-deferred practice from the original #4268 PR. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> --------- Co-authored-by: sim <sim@local> Co-authored-by: Claude Sonnet 5 <noreply@anthropic.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.