by mindfold-ai
The best agent harness.
# Add to your Claude Code skills
git clone https://github.com/mindfold-ai/Trellis| Capability | What it changes |
| --- | --- |
| Auto-injected specs | Write conventions once in .trellis/spec/, then let Trellis inject the relevant context into each session instead of repeating yourself. |
| Task-centered workflow | Keep PRDs, implementation context, review context, and task status in .trellis/tasks/ so AI work stays structured. |
| Parallel agent execution | Run multiple AI tasks side by side with git worktrees instead of turning one branch into a traffic jam. |
| Project memory | Journals in .trellis/workspace/ preserve what happened last time, so each new session starts with real context. |
| Team-shared standards | Specs live in the repo, so one person’s hard-won workflow or rule can benefit the whole team. |
| Multi-platform setup | Bring the same Trellis structure to 10 AI coding platforms instead of rebuilding your workflow per tool. |
# 1. Install Trellis
npm install -g @mindfoldhq/trellis@latest
# 2. Initialize in your repo
trellis init -u your-name
# 3. Or initialize with the platforms you actually use
trellis init --cursor --opencode --codex -u your-name
-u your-name creates .trellis/workspace/your-name/ for personal journals and session continuity.--cursor, --opencode, --iflow, --codex, --kilo, --kiro, --gemini, --antigravity, --qoder, and --codebuddy.No comments yet. Be the first to share your thoughts!
Put coding standards, file structure rules, review habits, and workflow preferences into Markdown specs. Trellis loads the relevant pieces automatically so you do not have to re-explain the repo every time.
Use git worktrees and Trellis task structure to split work cleanly across agents. Different tasks can move forward at the same time without stepping on each other’s branches or local state.
Task PRDs, checklists, and workspace journals make previous decisions available to the next session. Instead of starting from blank context, the next agent can pick up where the last one left off.
If your team uses more than one AI coding tool, Trellis gives you one shared structure for specs, tasks, and process. The platform-specific wiring changes, but the workflow stays recognizable.
Trellis keeps the core workflow in .trellis/ and generates the platform-specific entry points you need around it.
.trellis/
├── spec/ # Project standards, patterns, and guides
├── tasks/ # Task PRDs, context files, and status
├── workspace/ # Journals and developer-specific continuity
├── workflow.md # Shared workflow rules
└── scripts/ # Utilities that power the workflow
Depending on the platforms you enable, Trellis also creates tool-specific integration files such as .claude/, .cursor/, AGENTS.md, .agents/, .codex/, .kilocode/, and .kiro/. For Codex, Trellis now installs both project skills under .agents/skills/ and project-scoped config/custom agents under .codex/.
At a high level, the workflow is simple:
Specs ship as empty templates by default — they are meant to be customized for your project's stack and conventions. You can fill them from scratch, or start from a community template:
# Fetch templates from a custom registry
trellis init --registry https://github.com/your-org/your-spec-templates
Browse available templates and learn how to publish your own on the Spec Templates page.
--registry), parent-child subtasks, fix PreToolUse hook for CC v2.1.63+.trellis update, improved .gitignore handling, docs refresh./trellis:brainstorm.Those files are useful, but they tend to become monolithic. Trellis adds structure around them: layered specs, task context, workspace memory, and platform-aware workflow wiring.
No. Trellis currently supports Claude Code, Cursor, OpenCode, iFlow, Codex, Kilo, Kiro, Gemini CLI, and Antigravity. The detailed setup and entry command for each tool lives in the supported platforms guide.
No. Many teams start by letting AI draft specs from existing code and then tighten the important parts by hand. Trellis works best when you keep the high-signal rules explicit and versioned.
Yes. Personal workspace journals stay separate per developer, while shared specs and tasks stay in the repo where they can be reviewed and improved like any other project artifact.