import { GED_DIR } from "./contracts.js"; export interface StarterFile { path: string; content: string; } export const starterFiles: StarterFile[] = [ { path: `${GED_DIR}/VERSION`, content: `2 `, }, { path: `${GED_DIR}/CONTEXT-MAP.md`, content: `# Context Map Ged memory is current-state oriented. Durable root files describe the project as it is now; active work and runtime state live under branch/work scoped directories. ## Durable root memory - \`.ged/PROJECT.md\` — product goal, users, constraints, success criteria, repo signals. - \`.ged/ARCHITECTURE.md\` — current component boundaries and system shape. - \`.ged/PATTERNS.md\` — conventions and implementation patterns. - \`.ged/GLOSSARY.md\` — project/domain vocabulary. - \`.ged/DECISIONS.md\` — durable decisions and rationale. - \`.ged/STANDARDS.md\` — imported repo-wide agent standards. - \`.ged/SKILLS.md\` — durable skill guidance. ## Active work memory - \`.ged/work//SPEC.md\` - \`.ged/work//TASKS.md\` - \`.ged/work//TESTS.md\` - \`.ged/work//NOTES.md\` - \`.ged/work//META.json\` ## Runtime memory - \`.ged/runtime//STATE.md\` - \`.ged/runtime//SESSION-SUMMARY.md\` - \`.ged/runtime//checkpoints.json\` `, }, { path: `${GED_DIR}/PROJECT.md`, content: `# Project ## Goal Describe what this project should achieve. ## Users - Primary users: - Secondary users: ## Constraints - Technical constraints: - Product constraints: ## Success Criteria - What does success look like? `, }, { path: `${GED_DIR}/ARCHITECTURE.md`, content: `# Architecture Describe current system components, boundaries, and data flow. `, }, { path: `${GED_DIR}/PATTERNS.md`, content: `# Patterns Record implementation conventions and recurring workflow patterns. `, }, { path: `${GED_DIR}/GLOSSARY.md`, content: `# Glossary Record domain terms and definitions. `, }, { path: `${GED_DIR}/IDEAS.md`, content: `# Ideas ## Active ideas - ## Future ideas - ## Parking lot - `, }, { path: `${GED_DIR}/DECISIONS.md`, content: `# Decisions Record important choices here as the project evolves. ## Entries - Date: YYYY-MM-DD - Decision: - Why: - Impact: `, }, { path: `${GED_DIR}/STANDARDS.md`, content: `# Imported Standards These standards were imported from other harness-specific instruction files and approved for Ged use. No imported standards have been accepted yet. `, }, { path: `${GED_DIR}/SKILLS.md`, content: `# Skills ## Installed - None yet ## Recommended - None yet ## Deferred - None yet ## Rejected - None yet ## Usage Notes - Record why a skill was installed, recommended, or skipped. `, }, { path: `${GED_DIR}/project-skills/README.md`, content: `# Project Skills Store project-scoped skills that Ged auto-installs or creates for active tasks here. `, }, { path: `${GED_DIR}/SKILLS-STATE.json`, content: `{ "managed": [] } `, }, { path: `${GED_DIR}/work/root/SPEC.md`, content: `# Spec ## Problem ## Solution shape ## Key workflows ## Risks ## Open questions `, }, { path: `${GED_DIR}/work/root/TASKS.md`, content: `# Tasks ## Task slices | ID | Title | Depends On | Status | Done Criteria | | --- | --- | --- | --- | --- | `, }, { path: `${GED_DIR}/work/root/TESTS.md`, content: `# Tests ## Project-wide checks - ## Task-specific checks - ## Retry policy - Implementation retries before the plan must be tightened: 2 ## Recovery rule - If the same slice fails repeatedly, rewrite the slice, clarify the spec, and retry with a narrower plan. `, }, { path: `${GED_DIR}/work/root/NOTES.md`, content: `# Notes `, }, { path: `${GED_DIR}/work/root/META.json`, content: `{ "workId": "root", "schema": 1 } `, }, { path: `${GED_DIR}/PROGRESS.md`, content: `# Progress Ongoing log of project progress. `, }, { path: `${GED_DIR}/plans/INDEX.md`, content: `# Plan Index | ID | Title | Status | Created | Completed | | --- | --- | --- | --- | --- | `, }, { path: `${GED_DIR}/research/README.md`, content: `# Research Store external research summaries and package notes here. `, }, { path: `${GED_DIR}/specs/README.md`, content: `# Specs Store durable detailed specs here when they remain useful after active work completes. `, }, { path: `${GED_DIR}/tasks/README.md`, content: `# Task Artifacts Store per-task briefs, outputs, and failure histories here when they should outlive runtime summaries. `, }, { path: `${GED_DIR}/.gitignore`, content: `# Ephemeral session state runtime/ # Local repo-map cache artifacts REPO-MAP.md REPO-MAP.json `, }, { path: `.pi/agents/ged-brain.md`, content: `--- name: ged-brain description: GedPi brain for user-facing clarifying, planning, and implementation model: anthropic/claude-opus-4-6 tools: read, grep, find, ls, bash skill: ged-planning, ged-execution, ged-verification --- You are GedPi's only user-facing agent. Clarify with grill-me until the requested behavior, constraints, and success criteria are concrete enough to implement safely. Write durable project context into .ged/PROJECT.md and active work context into .ged/work//SPEC.md. Break the work into bounded slices in .ged/work//TASKS.md before editing code. Run the planned checks, record outcomes in .ged/runtime//STATE.md and .ged/runtime//SESSION-SUMMARY.md, and tighten the plan if a slice fails. `, }, ];