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/celestoai/agentor/agents-mdgit clone --depth 1 https://github.com/CelestoAI/agentorWhat 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 | $0.02517 | $0.02517 |
| Opus 5 | $0.01259 | $0.01259 |
| Sonnet 5 | $0.00503 | $0.00503 |
| Haiku 4.5 | $0.00252 | $0.00252 |
Grade C, and why
agentor AGENTS.md scanned grade C 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 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.
Hidden instructionshighPrompt injection
Directives inside HTML comments, invisible characters or bidirectional overrides are read by the model and not by the person reviewing the file.
<!-- This file is the single source of contributor guidance. .github/copilot-instructions.md points here rather than duplicating it: the two were kept "in sync" by hand, drifted anyway, and both went on describing subsys How it starts
The opening of the file, as written. The whole thing — 347 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Agentor
Project Overview
Agentor is an open-source framework for building AI agents with secure integrations across email, calendars, CRMs, and more. It connects LLMs to tools and services, and speaks both the Model Context Protocol (MCP) and Agent-to-Agent (A2A) protocols.
Since v0.1.0 it runs on its own agent engine; the openai-agents dependency is gone. See docs/dev/MIGRATION_PLAN.md for why.
Key Features:
- An agent loop the project owns end to end, emitting a typed event stream
- Durable runs: the event stream is persisted, so a run resumes after a crash
- Any OpenAI-compatible provider via
base_url, with litellm as the escape hatch - LiteMCP - FastAPI-compatible MCP server with decorator API
- A2A Protocol support for agent interoperability
- Tool registry and extensible tool system
Repository Structure
agentor/
├── src/agentor/ # Main package source code
│ ├── core/ # Agentor, the user-facing agent class
│ ├── engine/ # The agent loop: events, tools, models, store, tracing, mcp
│ ├── mcp/ # LiteMCP server, API router, proxy
│ ├── tools/ # Tool registry and implementations
│ ├── durable/ # Deprecation shim; durability moved into engine/store.py
│ ├── a2a.py # Agent-to-Agent protocol
│ ├── skills.py # Agent Skills loading
│ └── tool_search.py # Tool Search API
├── tests/ # Test suite
├── examples/ # Usage examples
├── docs/dev/ # Design and migration records
├── .github/ # GitHub configuration and workflows
└── pyproject.toml # Project metadata and dependencies
Development Setup
Prerequisites
- Python 3.11 or higher
- pip or uv package manager
Installation
- Clone the repository:
git clone https://github.com/CelestoAI/agentor.git
cd agentor
- Install dependencies using uv (recommended):
pip install uv
uv venv
uv sync
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 · 347 lines · 2,517 tokens per session scan C e90d4a70e864
agentor AGENTS.md is an instructions file published in the GitHub repository CelestoAI/agentor (192 stars, last pushed 8d ago), licensed Apache-2.0. It adds 2,517 tokens to every session, about $0.0126 per session on Opus 5. A static security scan graded it C with 1 finding (hidden instructions). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
visionagent CLAUDE.md
Instructions for sijeeshmiziha/visionagent, covering visionagent — claude code guide, project overview, architecture, key abstractions and development workflow.
graphiti AGENTS.md
Instructions for getzep/graphiti, covering repository guidelines, project structure & module organization, build, test, and development commands, coding style & naming conventions and testing guidelines.
autocontext CLAUDE.md
Instructions for greyhaven-ai/autocontext, covering claude.md, project overview, repository layout, commands and setup.
autocontext AGENTS.md
Instructions for greyhaven-ai/autocontext, covering agent guide, pick the right surface, working directories, setup and parity-last changes.
best-of-Agent-Harnesses CLAUDE.md
Claude Code instructions for RyanAlberts/best-of-Agent-Harnesses, covering repo notes for ai coding agents, commit identity (required), branching and pushing (required), verifying contribution credit (required) and how the list is generated.
linggen CLAUDE.md
Instructions for linggen/linggen, covering claude.md, doc and spec, build, test, run, architecture and configuration.