SLOPE

Sprint Lifecycle & Operational Performance Engine

Replace subjective retrospectives with quantified sprint metrics. Track every ticket's approach, outcome, and hazards — then use rolling analytics to spot patterns and improve over time.

Why SLOPE?

• Quantified retros — structured scorecards with objective scoring instead of "how did it feel?"
• Pattern detection — rolling handicap windows reveal if you consistently over-engineer, under-scope, or pick the wrong approach
• AI agent guidance — 16 guard hooks give real-time hints to Claude Code, Cursor, Windsurf, Cline, and OpenCode
• Pluggable metaphors — golf, tennis, baseball, gaming, D&D, matrix, or agile terminology — same math, your vocabulary
• Zero infrastructure — SQLite store, CLI-driven, lives in your repo

Quick Start

# Install
npm install -g @slope-dev/slope

# Initialize (auto-detects your AI coding tool)
slope init

# View your handicap card
slope card

Installed locally? Use npx slope instead.

Core Concepts

Every sprint has a par (expected baseline from ticket count: 1–2 → par 3, 3–4 → par 4, 5+ → par 5). Each ticket is a shot with an approach complexity (club: driver → putter) and an outcome (result: in_the_hole → missed). Over time, your handicap card shows rolling averages, miss patterns, and trend direction.

Features

Scoring & Analysis

• Scorecards — structured JSON retros with shot-by-shot tracking
• Handicap card — rolling windows (last 5, 10, all-time) with trend arrows
• Dispersion analysis — miss pattern heatmaps and area performance
• HTML reports — self-contained visual dashboards with charts
• Auto-card — generate draft scorecards from git commits + CI signals

Planning & Workflow

• Briefings — pre-sprint hazard index, nutrition alerts, filtered gotchas
• Sessions & claims — track who's working on what, detect conflicts
• Roadmap tools — validate dependencies, find critical path, parallel opportunities
• Club advisor — complexity recommendations based on historical performance

AI Agent Guidance

• 16 guard hooks — real-time hints injected into agent context
• MCP server — search API functions and execute SLOPE commands from your agent
• 5 platform adapters — Claude Code, Cursor, Windsurf, Cline, OpenCode
• Codebase map — auto-generated index for agent navigation

Team & Multi-Developer

• Team handicap — aggregate performance across team members
• Leaderboard — multi-developer performance ranking
• Standups — structured standup reports with handoff tracking
• Escalation — severity-based alerts for blocked work

Platform Compatibility

slope init --claude-code   # or --cursor, --windsurf, --cline, --opencode, --all

CLI Quick Reference

Setup & Config

Scoring

Planning

Sessions

Maintenance

Metaphors

7 built-in metaphors — same scoring math, different vocabulary:

slope init --metaphor=gaming          # Set during init
slope card --metaphor=tennis          # Override per-command

Documentation

• Getting Started — installation, setup, core concepts, platform guides
• Tutorial: First Sprint — end-to-end walkthrough with example scorecard
• Framework Reference — full scoring system specification
• Dashboard Guide — live performance dashboard
• Multi-Developer Guide — team handicap and leaderboard
• Cline Setup — Cline-specific configuration
• Scorecard Template — field-by-field reference

Core API

import {
  buildScorecard,
  validateScorecard,
  computeHandicapCard,
  computeDispersion,
  formatSprintReview,
  buildReportData,
  generateHtmlReport,
  loadScorecards,
  getMetaphor,
} from '@slope-dev/slope';

Published as @slope-dev/slope — includes the scoring engine, SQLite store, CLI (30 commands), and MCP server.

Contributing

1. Fork and clone
2. pnpm install && pnpm build
3. pnpm test — runs all tests
4. pnpm typecheck — type checking
5. Create a feature branch, make changes, submit a PR

License

MIT