Bootstrap GRACE framework structure for a new project. Use when starting a new project with GRACE methodology - creates docs/ directory, AGENTS.md, and XML templates for requirements, technology, development plan, verification plan, knowledge graph, and operational packet contracts.
Installs into .claude/skills of the current project.
Are you the author of Grace Init?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/osovv-grace-init)
---
name: grace-init
description: "Bootstrap GRACE framework structure for a new project. Use when starting a new project with GRACE methodology - creates docs/ directory, AGENTS.md, and XML templates for requirements, technology, development plan, verification plan, knowledge graph, and operational packet contracts."
---
Initialize GRACE framework structure for this project.
## Template Files
All documents MUST be created from template files located in this skill's `assets/` directory.
Read each template file, replace the `$PLACEHOLDER` variables with actual values gathered from the user, and write the result to the target project path.
| Template source | Target in project |
|------------------------------------------|-----------------------------|
| `assets/AGENTS.md.template` | `AGENTS.md` (project root) |
| `assets/docs/knowledge-graph.xml.template` | `docs/knowledge-graph.xml` |
| `assets/docs/requirements.xml.template` | `docs/requirements.xml` |
| `assets/docs/technology.xml.template` | `docs/technology.xml` |
| `assets/docs/development-plan.xml.template`| `docs/development-plan.xml` |
| `assets/docs/verification-plan.xml.template`| `docs/verification-plan.xml` |
| `assets/docs/operational-packets.xml.template`| `docs/operational-packets.xml` |
> **Important:** Never hardcode template content inline. Always read from the `.template` files — they are the single source of truth for document structure.
## Steps
1. **Gather project info from the user.** Ask for:
- Project name and short annotation
- Main keywords (for domain activation)
- Primary language, runtime, and framework (with versions)
- Key libraries/dependencies (if known)
- Testing stack (test runner, assertion style, mock/fake approach)
- Observability stack (logger, structured log fields, redaction constraints)
- High-level module list (if known)
- 2-5 critical flows or risky surfaces that must be verifiable early
2. **Create `docs/` directory and populate documents from templates:**
For each `assets/docs/*.xml.template` file:
- Read the template file
- Replace `$PLACEHOLDER` variables with user-provided values
- Write the result to the corresponding `docs/` path
3. **Create or verify `AGENTS.md` at project root:**
- If `AGENTS.md` does not exist — read `assets/AGENTS.md.template`, fill in `$KEYWORDS` and `$ANNOTATION`, and write to project root
- If `AGENTS.md` already exists — warn the user and ask whether to overwrite or keep the existing one
4. **Print a summary** of all created files and suggest the next step:
> "Run `$grace-plan` to design modules, data flows, and verification references. Then use `$grace-verification` to deepen tests, traces, and log-driven evidence before large execution waves. Use `docs/operational-packets.xml` as the canonical packet and delta reference during execution and refactors."
