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.
npx agentmods add instructions/agentsystemlabs/nebula/agents-mdgit clone --depth 1 https://github.com/AgentSystemLabs/nebulaWhat 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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.01263 | $0.01263 |
| Opus 5 | $0.00632 | $0.00632 |
| Sonnet 5 | $0.00253 | $0.00253 |
| Haiku 4.5 | $0.00126 | $0.00126 |
Grade B, and why
nebula AGENTS.md scanned grade B with 1 finding 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
grep when the index line is not enough (`grep -ril '<symbol or TERM>' .claude/memory/entries`). How it starts
The opening of the file, as written. The whole thing — 74 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Nebula
The shared memory and vocabulary
This repo runs a SELF-IMPROVING LOOP: every task starts from what past tasks recorded and ends by recording its own. Four committed files carry it. Read the index, the standing gotchas and the glossary before you start; the entries are fetched by index line or grep, never read wholesale.
| File | What it is | Cap |
|---|---|---|
.claude/MEMORY.md |
the index of the MEMORY LOG — one line per task: date, title, the TERMS and files it is about, its gotcha count, newest first | 200 lines |
.claude/memory/gotchas.md |
the standing gotchas — traps that outlive their task, one line each, grouped by TERM | 300 lines |
.claude/memory/entries/<date>-<slug>.md |
the entries — the full Asked / Did / Gotchas of each task | none |
TERMS.md |
the glossary — one ALL-CAPS name per feature, panel, key, command, hook route, daemon mechanism, status and workflow, with the user's words for it and where it lives | — |
.claude/memory/prs/ |
the PR ARCHIVE — every merged pull request rendered to Markdown (description, comments, reviews, inline comments) by a GitHub Action; not injected, reached by grep | — |
make ci enforces this: make memory-check (caps, index ↔ entries), make recall-eval (the
entries are still findable from their own prompts), make terms-check (no stale Where pointers or
dangling aliases).
Before you start a task
- Read
.claude/MEMORY.md,.claude/memory/gotchas.mdandTERMS.md. - Open the entries that match what the user is asking — same TERMS, crate, symptom or file;
grep when the index line is not enough (
grep -ril '<symbol or TERM>' .claude/memory/entries). Match on the user's vocabulary as well as your own: entries quote the original request verbatim.- a recorded gotcha is a mine already stepped on — do not step on it again
- a recorded decision ("we're not doing X because Y") is settled unless the user reopens it
- a recorded fix tells you where the code that matters actually lives
- all of it describes what was true when written: confirm a named file, function or flag still exists before relying on it, and correct it if it has gone stale
- Map every noun in the prompt onto
TERMS.mdthrough its Alias index ("top nav" → WORKSPACES BAR). A word that maps to two TERMS is an ambiguity to settle before working, not to guess. - Rewrite the prompt — PROMPT DADDY: read
.claude/skills/prompt-daddy/SKILL.mdand follow it. Do this on every prompt that is a task, before planning or grepping in earnest. The refined prompt is the request you work from. The skill lists what it skips (a reply to your own question, a bare confirmation, a specific mid-task correction, a skill trigger, and a pure question that changes nothing).
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.
- today Changed · +1 lines · +50 tokens per session 188d542be29f
- yesterday First seen · 73 lines · 1,213 tokens per session scan B 6bf187b0d3c4
nebula AGENTS.md is an instructions file published in the GitHub repository AgentSystemLabs/nebula (100 stars, last pushed today), licensed MIT. It adds 1,263 tokens to every session, about $0.0063 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
buildNext
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
next.js AGENTS.md
Instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
spec-kit AGENTS.md
Instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.
langchain AGENTS.md
Instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.