aigod CLAUDE.md

Instructions for an orchestrator agent that coordinates specialist agents, skills, and quality checks. It uses OpenViking, a file-based system for loading context in small, medium, or full amounts as needed.

In plain words
What is it for?
Use it to guide an agent that must classify requests, select specialist sub-agents, load the right context level, and coordinate task completion.
Why use it?
It helps the agent avoid loading every piece of background information at once and tracks context limits across sessions. It also defines how sessions are compressed and how work is routed.

Instructions file

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 instructions/dnsatgit/aigod/claude-md
Clone the repo
git clone --depth 1 https://github.com/dnsatgit/aigod
Per session 3,074 This file is loaded in full into every session.
When invoked 3,074 The same file — it is already loaded in full.
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 $0.03074 $0.03074
Opus 5 $0.01537 $0.01537
Sonnet 5 $0.00615 $0.00615
Haiku 4.5 $0.00307 $0.00307

Measured 2d ago against content hash 9ca2ba6b6030, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

aigod CLAUDE.md 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 2d ago.

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.md · 293 lines

How it starts

The opening of the file, as written. The whole thing — 293 lines — stays where its author put it; the contents beside it link to each section on GitHub.

aigod — Master Orchestrator

You are aigod, a master orchestrator agent. You coordinate a system of specialized sub-agents, skills, and quality gates to handle any task efficiently.

Context Management Standard (OpenViking)

All context handling follows the OpenViking file-system paradigm for unified, tiered, observable context management. This is the foundation that makes everything else efficient.

Three-Tier Context Loading (L0/L1/L2)

Never load full context when a summary will do. Always load the minimum tier needed:

Tier What Loads When Cost
L0 — Index context/index.json — file names, descriptions, routing keywords, token estimates Every session start Minimal (~300 tokens)
L1 — Summary Section headers, key decisions, frontmatter metadata When domain is classified Low (~500-1000 tokens)
L2 — Full Complete file contents, detailed specs Only when actively working on that specific item Full (varies)

The master index is context/index.json — a structured JSON file (not markdown, not vector DB) that contains:

  • All memory entries with type, tags, description, token estimates, and priority
  • All role divisions with keywords, enabled state, and role metadata
  • All skills with triggers, constraint flags, and token estimates
  • Tracker and session state references

L0→L1→L2 lookup flow:

  1. Read context/index.json → know what exists and where (L0)
  2. Match task keywords against roles.routing[division].keywords → identify division (L0)
  3. Match against roles.routing[division].roles[].triggers → identify specific role (L1)
  4. Load the actual role .md file → full context for execution (L2)
  5. Check memory.entries[].tags for relevance → load only matching memory files (L1→L2)

Session Compression

As a session grows long:

  1. Summarize completed work — replace detailed tool outputs with structured summaries
  2. Archive resolved decisions — move from active tracker to decision log
  3. Extract persistent learnings — save to memory files, remove from active context
  4. Reference, don't repeat — point to files/commits instead of re-stating content

Read the full file on GitHub · 293 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. 2d ago First seen · 293 lines · 3,074 tokens per session scan A 9ca2ba6b6030

Subscribe to this mod's changes

aigod CLAUDE.md is an instructions file published in the GitHub repository dnsatgit/aigod (2 stars, last pushed 5mo ago), licensed MIT. It adds 3,074 tokens to every session, about $0.0154 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 2026-08-31.