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/igptai/skills/claude-mdgit clone --depth 1 https://github.com/igptai/skillsWrote 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.
[](https://agentmods.dev/instructions/igptai/skills/claude-md)<a href="https://agentmods.dev/instructions/igptai/skills/claude-md"><img src="https://agentmods.dev/badge/instructions/igptai/skills/claude-md.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.03080 | $0.03080 |
| Opus 5 | $0.01540 | $0.01540 |
| Sonnet 5 | $0.00616 | $0.00616 |
| Haiku 4.5 | $0.00308 | $0.00308 |
Grade A, and why
skills 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 6d 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 — 261 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
Notes for any Claude session working in this repo. Keep this short. If something here is wrong, fix it — don't accumulate stale context.
What this is
A Claude Code plugin marketplace for iGPT. The marketplace manifest is at
.claude-plugin/marketplace.json. Each plugin under plugins/ bundles a
role-specific set of skills (sales, CS, finance, HR, etc.) that all
talk to the same backend MCP server: https://mcp.igpt.ai/. The marketplace
ships 13 plugins, 76 skills, and 13 agents (one per plugin) total at the
time of writing.
Working branch
dev. main is the published branch — don't commit there directly. PRs
from dev → main happen on demand, not yet wired up.
Conventions to follow
SKILL.md YAML frontmatter — single line for name and description,
plus a nested metadata block. No quotes:
---
name: skill-name-matches-directory
description: One-paragraph trigger description. Inline quotes for trigger phrases like "what's open" are fine — no escaping needed since the value isn't quoted.
metadata:
version: 1.0.0
---
Don't use folded multi-line descriptions (continuation lines are fragile;
one file got silently broken that way before). version is a string and
parses fine without quotes. Bump it when the skill's behavior changes.
.mcp.json in every plugin — exactly this shape, nothing more:
{
"mcpServers": {
"igpt": {
"type": "http",
"url": "https://mcp.igpt.ai/"
}
}
}
No description field. type is http, not url.
Prerequisites section in SKILL.md — inline, not a remote fetch. Every
SKILL.md has a ## Prerequisites block right after the H1 that tells Claude
what to do if the iGPT MCP isn't connected or auth fails. The block is
self-contained so a user who installs a single skill outside the plugin still
gets useful guidance. The full guard at shared/mcp-guard.md (canonical URL
in each block) is kept as a fallback for deeper troubleshooting only — not
fetched on every invocation.
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.
- 6d ago First seen · 261 lines · 3,080 tokens per session scan A a19b3e035880
skills CLAUDE.md is an instructions file published in the GitHub repository igptai/skills (16 stars, last pushed 4mo ago), licensed MIT. It adds 3,080 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 2026-08-30.
Other instructions, from other repositories
langgraph AGENTS.md
AGENTS.md instructions for langchain-ai/langgraph, covering agents instructions, corridor security analysis, libraries and dependency map.
hatch3r CLAUDE.md
Claude Code instructions for hatch3r/hatch3r, covering hatch3r — development instructions, architecture, development commands, two-axis pillar framework (2.0.0) and orchestrator self-discipline (bypass protection).
aeon CLAUDE.md
Claude Code instructions for aeonfun/aeon, covering aeon, how aeon works, strategy, voice and soul file hierarchy (read in this order).
cortex AGENTS.md
AGENTS.md instructions for tools-for-agents/cortex, covering agents.md — cortex, setup, the rules this repo is built on, tests and ci.
nextjs-skills CLAUDE.md
Instructions for mohamed-hossam1/nextjs-skills: All contribution rules — style, naming, file layout, verification — live in AGENTS.md. Read that first.
nocturnusai CLAUDE.md
Claude Code instructions for Auctalis/nocturnusai, covering claude.md, project overview, build & run commands, build all modules and run the http api server (port 9300).