889c7efba0
Phase 3 of epic #4636. Tests only; no implementation. These MUST fail. A refactor changes what a good test looks like: the behavior under test must be IDENTICAL before and after, so most of this phase's safety comes from invariance rather than new assertions. That safety net already exists and is untouched here — tests/security.test.cjs already pins the two engine behaviors a re-derivation would silently lose: :177 a DANGLING symlink to a non-existent OUTSIDE target stays safe:false (the existence-oracle closure) :213 a not-yet-created file in a not-yet-created subdir under a non-canonical base stays safe:true (ancestor canonicalization) plus traversal, absolute in/out, null bytes, empty, non-string, and requireSafePath's throw. Those 0 deletions are the point: if any of them had to change, the engine would have changed, and the engine is not supposed to. What is new is the export surface Phase 3 introduces: assertWithinRoot(candidate, root, label?, opts?) -> ContainedPath (throws) tryWithinRoot(candidate, root, opts?) -> ContainedPath | null Two shapes rather than one, because several call sites need a NON-throwing check — findPhaseArtifact probes a direct path, then a .planning/ path, then each readdir entry, and throwing on the first miss would break it outright. ADR-4650 names only the throwing form; this is the gap between the ADR and the call sites, recorded rather than papered over. Rows that exist because they are the ones nobody enumerates: - tryWithinRoot must return EXACTLY null on escape, and its return must not contain the escaping path's basename. The shape being replaced populates its "resolved" field with the escaping path precisely on the traversal branch, so a caller who ignores the boolean gets a usable attacker-controlled value. That is the defect the narrowing exists to remove, so it is asserted directly. - A seeded parity property: tryWithinRoot returns non-null if and only if assertWithinRoot does not throw, and the values agree. Two exported shapes over one engine is a divergence pair by construction. - The rejection text still contains the phrase "escapes allowed directory". Another suite surfaces it through a user-facing "reason" field, and a refactor is exactly where wording drifts unnoticed. Co-Authored-By: Claude Opus 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.