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/swoopeagle/standardgraph/agents-mdgit clone --depth 1 https://github.com/swoopeagle/standardgraphWhat 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 | $0.02148 | $0.02148 |
| Opus 5 | $0.01074 | $0.01074 |
| Sonnet 5 | $0.00430 | $0.00430 |
| Haiku 4.5 | $0.00215 | $0.00215 |
Grade A, and why
standardgraph AGENTS.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 3d 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.
How it starts
The opening of the file, as written. The whole thing — 166 lines — stays where its author put it; the contents beside it link to each section on GitHub.
StandardGraph — Codex context
What this is
FastMCP server exposing 175,000+ education standards across 310 curriculum systems as six MCP tools for Codex Desktop. Standards cover Math, Science, ELA, Social Studies, CS, Arts, and World Languages.
Architecture
packages/
common-core/ → PyPI package "standardgraph" — the MCP server
src/common_core/
server.py → all six MCP tools (search, lookup, progression, learning_path, map, list)
config.py → DB_PATH resolution (~/.standardgraph/common_core.db)
ingestion/ → pipeline: fetchers → embed → relate → crosswalk
shared/ → shared DB helpers
data/common_core.db → dev/pipeline DB (used by overnight_run.sh)
~/.standardgraph/common_core.db → installed user DB (used by MCP server)
scripts/
mcp_test.py → 333-test suite (imports server directly, no MCP protocol)
overnight_run.sh → full ingestion pipeline (run on Mac Studio overnight)
dashboard.sh → hardware + pipeline progress dashboard
progress.sh → pipeline-only progress view
Key facts
- DB size: ~1.9 GB
- Standards: 175,738 across 310 systems (incl. CCSS sub-standard decomposition, source-side decomposition of 11 high-bundling systems, CCSS Mathematical Practice standards, and the 2026-07 international expansion incl. 10 African systems)
- Crosswalk rows: ~117,699 (hub-centric: CCSS for math, NGSS for science, etc.)
- Crosswalk quality scores: ~75,055 rows (~63.8%) carry a 1–5 quality score (LLM rubric scoring + deterministic exact-match); pre-existing AP/IB source rows are scored. Remainder unscored (
nlp_pass, ranked by cosine, treated as neutral quality) — includes new math mappings added by the 2026-07 decomposition/MP regeneration. - Relationships: ~3.79M rows (prerequisites/successors)
- Ollama host:
http://169.254.1.1:11434(Mac Studio via Thunderbolt Bridge from Mini 2 — 0.4ms RTT) - HuggingFace dataset:
swoopeagle/standardgraph(file:common_core.db) - PyPI package:
standardgraph
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.
- 3d ago First seen · 166 lines · 2,148 tokens per session scan A c72e03f178d7
standardgraph AGENTS.md is an instructions file published in the GitHub repository swoopeagle/standardgraph (5 stars, last pushed 1mo ago), licensed MIT. It adds 2,148 tokens to every session, about $0.0107 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.
Other instructions, from other repositories
exam-revision-handbook AGENTS.md
Instructions for mianbaofang/exam-revision-handbook, covering project agent guidance, start here, single canonical skill, product boundaries and change discipline.
hult-cohort-program CLAUDE.md
Instructions for rogerSuperBuilderAlpha/hult-cohort-program: Read AGENTS.md — this repository is the Hult Cohort Program monorepo (curriculum + platform).
AI-learning-by-claude-code CLAUDE.md
Claude Code instructions for CyrusZhang23/AI-learning-by-claude-code, covering ai-learning-by-claude-code — teaching interface, teaching protocol (most important — always follow it), course scope (locked), course index and command interface.
AI-learning-by-claude-code AGENTS.md
AGENTS.md instructions for CyrusZhang23/AI-learning-by-claude-code, covering ai-learning-by-claude-code — codex entry point, 给 codex 的首要指令（每次会话开始执行）, codex 用户怎么开课, 命令语法差异（速查） and codex 教学执行规则.
open-dictionary AGENTS.md
Instructions for ahpxex/open-dictionary, covering open dictionary rewrite charter, product framing, core workflow, technical framework and 1. raw ingestion layer.
designing-real-world-ai-agents-workshop CLAUDE.md
Instructions for iusztinpaul/designing-real-world-ai-agents-workshop, covering project, project structure, tech stack, access documentation and running qa.