edpa

EDPA — Evidence-Driven Proportional Allocation

Derive hours from Git evidence. No timesheets.

Score = JobSize x ContributionWeight x RelevanceSignal
DerivedHours = (Score / SumScores) x Capacity
Guarantee: Sum(DerivedHours) = Capacity (always)

The Problem

Your team spends hours filling timesheets. The data is inaccurate, the process is hated, and for audit-grade projects (EU grants, government contracts) it's a compliance nightmare.

The Solution

EDPA eliminates manual timesheets entirely. Your team works normally — commits, PRs, reviews, comments — and EDPA derives hours automatically from this delivery evidence.

Before EDPA:

Monday morning: "What did I work on last week? Let me guess... 4h on S-200, maybe 6h on F-102..."

After EDPA:

$ python3 .claude/edpa/scripts/engine.py --edpa-root .edpa --iteration PI-2026-1.3

EDPA 1.18.5 — Iteration PI-2026-1.3
======================================================================
Person                    Role     Capacity  Derived  Items   OK
----------------------------------------------------------------------
J. Urbanek                Arch          40h    40.0h     15   OK   ← Arch credited 15×
                                                                     for Feature/Epic gate
                                                                     transitions (LBC, design,
                                                                     refinement) — invisible
                                                                     in old simple mode.
O. Tuma                   DevSecOps      80h    80.0h      9   OK
Turyna                    Dev           60h    60.0h      7   OK
Matousek                  Dev           60h    60.0h      5   OK
----------------------------------------------------------------------
TEAM TOTAL                             240h   240.0h
PLANNING CAPACITY                    192.0h  (factor: 0.8)

All invariants passed: YES

Key Features

• Zero manual input — hours derived from GitHub delivery evidence (commits, PRs, reviews, comments)
• Mathematical guarantee — derived hours always sum to declared capacity
• Gates mode (default) — credits each Initiative/Epic/Feature status transition as a mini-deliverable, so prep work (LBC, decomposition, design) gets credited as it happens, not only at final Done. Validated to ±0.35 pp stability under ±20 % CW perturbation across 100 Monte Carlo runs.
• Bidirectional GitHub Projects sync — sync push creates issues with custom fields, sync pull mirrors GH UI changes back into local YAML; conflict auto-resolution with last-write-wins / local-wins / remote-wins strategies.
• Dual-view — per-person timesheets AND per-item cost allocation from the same data
• Audit-grade — frozen snapshots, immutable records, BankID signing support
• Self-tuning — auto-calibrates heuristics using Karpathy's autoresearch loop
• GitHub-native — works with GitHub Issues, Projects, PRs, and Actions (real E2E tested against technomaton/edpa-e2e-test sandbox)

First 5 minutes — guided walkthrough

This walkthrough takes a fresh empty repo to a closed iteration with derived hours and per-person reports. No GitHub Project required — the walkthrough stays local so onboarding is zero-friction. For the real GitHub Projects flow (push backlog, sync issue states, gate-based prep-work attribution) see docs/RUNBOOK.md.

1. Install (~30 s)

mkdir my-edpa-toy && cd my-edpa-toy
git init -q
curl -fsSL https://edpa.technomaton.com/install.sh | sh

You should see:

EDPA Installer
  Python 3.X ✓
  PyYAML ✓
  mcp (MCP SDK) ✓
  openpyxl ✓
  ...
EDPA 1.18.5 installed successfully!

Three config files were seeded from templates:

ls .edpa/config/
# edpa.yaml  heuristics.yaml  people.yaml

2. Edit people.yaml to your team (~1 min)

The template ships with placeholder names. Replace them with your team — even one person works. Minimum example for the AI-native default (1-week iterations, 5-week PI = 4 delivery + 1 IP):

cadence:
  iteration_weeks: 1        # AI-native default; use 2 for classic SAFe
  pi_weeks: 5               # 4 delivery iterations + 1 IP

people:
  - id: alice
    name: "Alice Architect"
    role: Arch
    fte: 0.5
    capacity_per_iteration: 20    # FTE × 40 for 1-week iter
  - id: bob
    name: "Bob Developer"
    role: Dev
    fte: 1.0
    capacity_per_iteration: 40

Verify the engine sees them:

python3 .claude/edpa/scripts/engine.py --status