gan

gan is a command for coding agents from datacore-one/datacore. It costs 36 tokens per session (536 once invoked), scanned A, original, MIT.

Adversarial multi-agent build loop. Plan → Generate → Evaluate → iterate. Pairs with /office-hours for ideation and /plan-ceo-review for scope.

Command

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/datacore-one/datacore/gan
Clone the repo
git clone --depth 1 https://github.com/datacore-one/datacore

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for gan

README.md
[![agentmods](https://agentmods.dev/badge/commands/datacore-one/datacore/gan.svg)](https://agentmods.dev/commands/datacore-one/datacore/gan)
Your own site
<a href="https://agentmods.dev/commands/datacore-one/datacore/gan"><img src="https://agentmods.dev/badge/commands/datacore-one/datacore/gan.svg" alt="Measured on agentmods" height="20"></a>
Per session 36 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 536 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin unknown No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00036 $0.00536
Opus 5 $0.00018 $0.00268
Sonnet 5 $0.00007 $0.00107
Haiku 4.5 $0.00004 $0.00054

Measured today against content hash 373238ea498b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

gan scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured today.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.datacore/commands/gan.md · 80 lines

What it actually says

/gan Command

Build products through adversarial multi-agent iteration.

Usage

/gan <one-line description>     # Start from scratch
/gan --spec <path>              # Start from existing spec
/gan --from-office-hours        # Continue from last /office-hours output

For best results, chain with gstack skills:

/office-hours          → Brainstorm & validate the idea (forcing questions)
  ↓
/plan-ceo-review       → Expand scope, find the 10-star product
  ↓
/gan                   → Adversarial build: Plan → Generate → Evaluate → iterate
  ↓
/plan-eng-review       → Lock in architecture before shipping
  ↓
/qa                    → Systematic QA testing

You can enter at any stage. /gan works standalone for well-defined briefs.

How It Works

Step 1: Plan (Opus)

  • Expand brief into full product spec
  • Features, design direction, evaluation rubric
  • Present to user for approval/modification

Step 2: Generate (Sonnet)

  • Implement the spec
  • Follow technical stack and design direction exactly
  • On subsequent iterations, focus ONLY on evaluator feedback

Step 3: Evaluate (Opus)

  • Test against rubric (functionality, design, code quality, UX, performance)
  • Score each criterion 0-10
  • List specific, actionable fixes with file:line references

Step 4: Decision Gate

  • All criteria >= 8/10 → DONE, present to user
  • Any criterion < 8/10 → Feed feedback to Generator, iterate
  • Max 3 iterations → Present best attempt with evaluator notes

Integration with Forge

For Forge product generation:

/gan "Etsy listing for [product idea]"

The GAN harness is particularly powerful for Forge because:

  • Planner generates product spec with marketplace positioning
  • Generator creates the product assets/listing
  • Evaluator checks against marketplace best practices

Agent

Coordinator: gan-harness agent Subagents: Planner (Opus), Generator (Sonnet), Evaluator (Opus)

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. today First seen · 80 lines · 36 tokens per session scan A 373238ea498b

Subscribe to this mod's changes

gan is a command published in the GitHub repository datacore-one/datacore (4 stars, last pushed today), licensed MIT. It adds 36 tokens to every session and 536 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.