Status: pre-launch. This page describes target design; this document does not establish a deployed service.

C0 -- Developer Quick Start

Legacy dev snapshot; use only inside an active implementation sprint after current environment verification.

Prerequisites

• Python 3.11+
• Git
• An OpenAI API key (or Anthropic, or Google -- any supported provider)

Steps

# 1. Clone the organism
git clone https://github.com/skyzai/organism.git
cd organism

# 2. Navigate to the agent pipeline (private path, not reproduced here)

# 3. Install dependencies
pip install openai sentence-transformers

# 4. Set your API key
export OPENAI_API_KEY=sk-your-key-here

# 5. Run a test signal through the pipeline
python3 -m pipeline.run \
  --provider openai \
  --model gpt-4o \
  "Test signal: BTC whale moves 10k to exchange"

What Happens

Your signal moves through the pipeline:

Signal received
  -> Circle adapter (IS): classify, verify, score
  -> ReFu adapter (COULD): probability assessment
  -> APU adapter (SHOULD): Council deliberation
-> Executor (ACT): unsigned ProposalEnvelope generated

Output: unsigned ProposalEnvelope with PROCEED/HOLD recommendation; authority remains external

You should see structured JSON output with the Council's recommendation within 10-20 seconds due to be finalized through implementation evidence on your model choice.

Alternative Providers

# Anthropic
export ANTHROPIC_API_KEY=sk-ant-your-key-here
python3 -m pipeline.run --provider anthropic --model claude-sonnet-4-20250514 "Test signal"

# Google
export GOOGLE_API_KEY=your-key-here
python3 -m pipeline.run --provider google --model gemini-2.5-pro "Test signal"

# Local (SmallEBM -- no API key needed)
python3 -m pipeline.run --provider local --model small-ebm "Test signal"

Verify the Installation

python3 -m pipeline.heartbeat
# [OK] Config loaded  [OK] Provider: openai  [OK] Schemas valid  [OK] Heartbeat: alive

Next Steps

• C1 Pipeline Architecture -- the full 13-file breakdown
• C2 Adapters -- inter-organ wiring
• C5 Uplink -- complete orientation and route into the current uplink spine

Troubleshooting

Problem	Fix
ModuleNotFoundError	Ensure you are in the agents/ directory
AuthenticationError	Check your API key export
TimeoutError	Model provider may be rate-limiting you
Empty output	Try --verbose flag for debug logging

---

Implementation Acceptance Checklist

Before calling your local setup “ready,” confirm all of the following:

• Repository clones without manual patching
• Dependency install completes without lockfile drift
• Core pipeline command runs start-to-finish at least once
• Build artifacts are reproducible on a second machine/user account
• Environment variable contract is documented in a local .env.example

Common Failure Modes

1. Version skew (Node/Python toolchain mismatch)
2. Hidden local state (cache creates false green)
3. Undeclared env vars (works for one operator only)
4. Path assumptions (relative path breakage in scripts)

Minimum Independently Verified Handoff Standard

If a second operator can go from clone -> run -> output in <30 minutes without ad-hoc Slack help, this page is operationally complete.

---