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/zzhiyuann/cortex/claude-mdgit clone --depth 1 https://github.com/zzhiyuann/cortexWrote 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/zzhiyuann/cortex/claude-md)<a href="https://agentmods.dev/instructions/zzhiyuann/cortex/claude-md"><img src="https://agentmods.dev/badge/instructions/zzhiyuann/cortex/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 | $0.01063 | $0.01063 |
| Opus 5 | $0.00531 | $0.00531 |
| Sonnet 5 | $0.00213 | $0.00213 |
| Haiku 4.5 | $0.00106 | $0.00106 |
Grade A, and why
cortex CLAUDE.md scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -s -X POST "https://api.telegram.org/bot$BOT_TOKEN/sendMessage" \ How it starts
The opening of the file, as written. The whole thing — 107 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Cortex — Autonomous Improvement Agent
Mission
You are the autonomous improvement agent for the Cortex ecosystem. You act as the founder's proxy — making product, UX, and engineering decisions to ship meaningful improvements. Work systematically, commit frequently, and send Telegram updates at milestones.
Language
- All code, comments, commit messages, docstrings: English
- Telegram messages to the user: Chinese
Communication
Send Telegram updates (ONLY at real milestones, not every step).
Read credentials from ~/.cortex/config.yaml (generated by cortex init):
BOT_TOKEN=$(python3 -c "import yaml; c=yaml.safe_load(open('$HOME/.cortex/config.yaml')); print(c['telegram']['bot_token'])")
CHAT_ID=$(python3 -c "import yaml; c=yaml.safe_load(open('$HOME/.cortex/config.yaml')); print(c['telegram']['chat_id'])")
curl -s -X POST "https://api.telegram.org/bot$BOT_TOKEN/sendMessage" \
-H 'Content-Type: application/json' \
-d "$(python3 -c "import json; print(json.dumps({'chat_id': $CHAT_ID, 'text': '你的消息'}))")"
When to send:
- Session started (what you plan to work on)
- Major feature completed (what changed, how to test)
- Stuck on something (what, why, what you need)
- Session ending (summary of all changes)
Git
Auto commit + push after each meaningful change. Do NOT batch everything into one giant commit. Each commit should be one logical improvement.
Monorepo Structure
This is a uv workspace monorepo. All packages live under packages/:
cortex/ # repo root
├── packages/
│ ├── cortex-cli/ # Umbrella CLI
│ ├── vibe-replay/ # Session capture & visualization (PRIMARY FOCUS)
│ ├── forge/ # Tool generation agent
│ ├── a2a-hub/ # Agent communication hub
│ └── dispatcher/ # Telegram agent bridge
├── pyproject.toml # workspace definition
├── index.html # GitHub Pages website
└── CLAUDE.md # this file
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.
- 2d ago First seen · 107 lines · 1,063 tokens per session scan A 1086afb40a65
cortex CLAUDE.md is an instructions file published in the GitHub repository zzhiyuann/cortex (0 stars, last pushed 5mo ago), licensed MIT. It adds 1,063 tokens to every session, about $0.0053 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
dibs AGENTS.md
Instructions for Agenxy/dibs, covering agents.md: orientation for agents working on dibs, what this repo is, layout, rules you must not break and working here.
SuperAgenticMCP AGENTS.md
AGENTS.md instructions for ANAMIZED/SuperAgenticMCP, covering agents.md — superagenticmcp, what this project is, how to run & verify, hard rules for agents and surfaces that must stay working.
agent-handoff-mcp AGENTS.md
Instructions for SarutobiSasuke8/agent-handoff-mcp: This repository is the public, vendor-neutral Agent Handoff MCP engine.
dibs CLAUDE.md
Instructions for Agenxy/dibs, a project described as: Keeps your agents in the loop about each other. See what the fleet is doing, then message, sync and send files between agents. Dibs reports; it never acts.
apra-fleet AGENTS.md
AGENTS.md instructions for Apra-Labs/apra-fleet, covering apra fleet - agent context, dev commands, conventions, deepwiki and beads issue tracker.
apra-fleet CLAUDE.md
Claude Code instructions for Apra-Labs/apra-fleet, covering apra fleet - claude code context, dev commands, conventions, deepwiki and beads issue tracker.