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/jokull/icelandic-data/agents-mdgit clone --depth 1 https://github.com/jokull/icelandic-dataWhat 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 | $0.06945 | $0.06945 |
| Opus 5 | $0.03472 | $0.03472 |
| Sonnet 5 | $0.01389 | $0.01389 |
| Haiku 4.5 | $0.00694 | $0.00694 |
Grade A, and why
icelandic-data 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 — 520 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Icelandic Data Toolkit
Data toolkit for Icelandic public data. Each skill in .agents/skills/ documents one data source — API endpoints, series codes, encoding quirks, classification changes. Scripts in scripts/ fetch, clean, and transform the data.
Architecture
/.agents/skills/{source}/SKILL.md # Data source docs (API, endpoints, caveats)
/scripts/{source}.py # Fetch + transform scripts (uv project)
/tests/health/test_{source}.py # Upstream health probe (marker: health)
/data/
/raw/{source}/ # Raw downloads (Excel, CSV, JSON)
/processed/ # Cleaned datasets
Skills follow the agentskills.io open standard, so the same
SKILL.md files work in both Claude Code and Codex:
.agents/skills/is the real location — read natively by Codex..claude/skillsis a symlink to it, so Claude Code discovers the same files.CLAUDE.mdis a symlink toAGENTS.md, so both agents read these instructions.
Skill names must be lowercase, ASCII, hyphen-separated (no underscores, no ð/æ).
Skills
Each skill documents ONE data source:
- API endpoints and authentication
- Available series and their scope
- Tariff codes, variable mappings, classification changes
- Example fetch commands
- Known caveats (encoding, date ranges, schema changes)
There is deliberately no index of skills here. Each SKILL.md carries a description
in its frontmatter, and that description is what both agents preload and match against —
a table in this file would be a second, staler copy that cannot trigger anything.
Run ls .agents/skills/ to enumerate them.
Keep each description under ~160 characters: Codex truncates when all descriptions
combined exceed 8,000 characters, and with 45 skills that budget is the binding constraint.
When asked about a new data source: follow the new-data-source skill.
HTML Reports
When asked for a report, produce a single self-contained .html file in /reports/:
- Embed data as JSON in
<script>tags - Use Chart.js (CDN) —
<script src="https://cdn.jsdelivr.net/npm/chart.js@4"></script> - No build step, no dependencies
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 · 520 lines · 6,945 tokens per session scan A f5d48eef7a74
icelandic-data AGENTS.md is an instructions file published in the GitHub repository jokull/icelandic-data (52 stars, last pushed 3d ago), licensed MIT. It adds 6,945 tokens to every session, about $0.0347 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 2026-08-30.
Other instructions, from other repositories
GPT-RAG release.instructions.md
Instructions for Azure/GPT-RAG, a project described as: Sharing the learning along the way we been gathering to enable Azure OpenAI at enterprise scale in a secure manner. GPT-RAG core is a Retrieval-Augmented Generation pattern running in Azure, using Azure Cognitive Search for retrieval and Azure OpenAI large…
OpenCandle AGENTS.md
Instructions for Kahtaf/OpenCandle, covering opencandle, commands, where to look, code style and conventions.
oci-agent CLAUDE.md
Instructions for Netflix-Skunkworks/oci-agent, covering observational causal inference (oci) agent and rules.
ZipAI CLAUDE.md
Claude Code instructions for nickdesi/ZipAI, covering claude.md — zipai: ultra-dense token optimizer, rules, 1. zero filler, 2. ambiguity and 3. prompt caching.
TreeSkill CLAUDE.md
Instructions for JimmyMa99/TreeSkill, covering claude.md, project overview, commands, install and run tests.
AutoRAG-Research CLAUDE.md
Instructions for NomaDamas/AutoRAG-Research, covering claude.md, project overview, common commands, setup and code quality.