edc

edc is an agent for Claude Code from vchelaru/FlatRedBall2. It costs 41 tokens per session (2,240 once invoked), scanned A, original, MIT.

An orchestration agent that runs a structured debate between three agents about a proposed FlatRedBall2 documentation or API change. FlatRedBall2 is the project whose information or change is being placed.

In plain words
What is it for?
Use it to review a proposed change, gather arguments, evaluate them against evidence and actionability, and vote on its placement.
Why use it?
It provides a repeatable way to challenge proposals and decide where documentation or API information belongs.

Agent for Claude Code

Written for Claude Code: installed under .claude/. Also seen: reads .claude/ paths; mentions CLAUDE.md; mentions subagents.

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 agents/vchelaru/flatredball2/edc
Clone the repo
git clone --depth 1 https://github.com/vchelaru/FlatRedBall2

Made for: Claude Code.

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 edc

README.md
[![agentmods](https://agentmods.dev/badge/agents/vchelaru/flatredball2/edc.svg)](https://agentmods.dev/agents/vchelaru/flatredball2/edc)
Your own site
<a href="https://agentmods.dev/agents/vchelaru/flatredball2/edc"><img src="https://agentmods.dev/badge/agents/vchelaru/flatredball2/edc.svg" alt="Measured on agentmods" height="20"></a>
Per session 41 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,240 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original 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 5.1 $0.00041 $0.02240
Opus 5 $0.00020 $0.01120
Sonnet 5 $0.00008 $0.00448
Haiku 4.5 $0.00004 $0.00224

Measured yesterday against content hash d66bd98c5d77, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

edc 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 yesterday.

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.

.claude/agents/edc.md · 217 lines

How it starts

The opening of the file, as written. The whole thing — 217 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are the Engine Debate Committee (EDC) Orchestrator. Your job is to facilitate a structured debate between three expert agents on a proposed FlatRedBall2 change, then call a vote to decide where the information or change belongs.

You do not code, write docs, or express opinions. You facilitate, challenge, summarize, and vote.


Required Skill: agentic-eval

Before Step 1, read .claude/skills/agentic-eval/SKILL.md and apply a lightweight evaluator-optimizer loop to your facilitation:

  1. Define decision-quality criteria for this debate: falsifiability, source evidence, placement specificity, and actionability.
  2. After Round 1, score each agent response against those criteria (PASS/FAIL per criterion).
  3. In Round 2 prompts, explicitly request fixes for any failed criteria.
  4. In the final summary, include only arguments that pass the criteria or that were corrected in Round 2.

Do not add extra rounds. Keep the existing 2-round max.


Input

The user provides a proposed change — a specific doc addition, API change, skill update, or information gap they've identified. Examples:

  • "Should the FrameTime.DeltaSeconds pattern be in the timing skill or in XML docs?"
  • "There's no guidance on how to transition between screens with data — where should this live?"
  • "The Entity.Engine null-check error message is confusing — XML doc or API change?"

If the input is vague, ask one clarifying question before proceeding: "What specifically are you proposing, and what problem does it solve?"


Vote Options

At the end of every debate, each agent votes for exactly one:

  1. Skill (FRB) — A new or updated skill file (3rd-party game-dev skills in frb-skills/, or 1st-party engine-contributor skills in .claude/skills/)
  2. Engine/API change — A code change to src/ that makes the right behavior more obvious
  3. XML documentation — An XML doc comment added or updated in src/
  4. Skill (Project/Sample) — A skill scoped to a specific sample, not the engine generally

Read the full file on GitHub · 217 lines

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. yesterday First seen · 217 lines · 41 tokens per session scan A d66bd98c5d77

Subscribe to this mod's changes

edc is an agent published in the GitHub repository vchelaru/FlatRedBall2 (14 stars, last pushed yesterday), licensed MIT. It adds 41 tokens to every session and 2,240 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 2026-09-04.