e2a AGENTS.md

Project instructions for e2a, a service whose public repository must not contain customer or production-derived data. They define the repository layout, documentation sources, and development rules.

In plain words
What is it for?
Use them when changing, testing, documenting, or reviewing this public repository. They are especially relevant when handling customer identifiers, messages, URLs, timestamps, or generated output.
Why use it?
They make the boundary between safe public examples and private customer information explicit. This helps prevent accidental disclosure in code, tests, issues, logs, screenshots, or commits.

Instructions file for CodexOpenCode

Install

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.

agentmods
npx agentmods add instructions/tokencanopy/e2a/agents-md
Clone the repo
git clone --depth 1 https://github.com/tokencanopy/e2a

Made for: Codex, OpenCode.

Per session 6,823 This file is loaded in full into every session.
When invoked 6,823 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.06823 $0.06823
Opus 5 $0.03411 $0.03411
Sonnet 5 $0.01365 $0.01365
Haiku 4.5 $0.00682 $0.00682

Measured yesterday against content hash d0e192467c73, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

e2a 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 yesterday.

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.

AGENTS.md · 449 lines

How it starts

The opening of the file, as written. The whole thing — 449 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md

Guidance for AI coding agents working in this repository (Claude Code reads this file via the @AGENTS.md import in CLAUDE.md). Assumes no prior knowledge of the project. For deeper prose, see README.md (product), CONTRIBUTING.md (contributor workflow), docs/ (API reference, deployment, design docs), and SECURITY.md.

Public repository data boundary (non-negotiable)

This repository and all of its GitHub surfaces are public. Never put real customer data or non-public production-derived data in source files, tests, fixtures, generated artifacts, commit messages, branch names, PR/issue/discussion titles or bodies, review comments, screenshots, log excerpts, or benchmark output. This includes email and agent addresses, customer domains or names, user/account/agent/message/contact IDs, message content or subjects, customer-specific URLs or IP addresses, timestamps tied to customer activity, and delivery, bounce, or incident details that can be correlated to a customer. An address on agents.e2a.dev is a customer identifier, not a safe example value. Public service metadata deliberately documented for users, such as https://api.e2a.dev, is not customer data.

Before any GitHub-facing write, replace production-derived values with synthetic ones such as example.com, .test, .invalid, agents.localhost, or clearly fictional resource IDs. Describe production evidence only in anonymized, non-correlatable terms (for example, "a production customer saw repeated SMTP rejections"). Keep exact evidence only in an approved private incident or operations system. Relevance to a fix is not an exception to this boundary.

If customer data is found before push, stop and rewrite the affected files and commits. If it reaches GitHub, treat it as a data exposure: stop propagation, notify Josh/security, sanitize the PR or issue, rewrite reachable Git history, check related refs/forks/caches, and contact GitHub Support when cached views or references need purging. Deleting the visible text alone is not sufficient.

Read the full file on GitHub · 449 lines

Changes

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.

  1. yesterday First seen · 449 lines · 6,823 tokens per session scan A d0e192467c73

Subscribe to this mod's changes

e2a AGENTS.md is an instructions file published in the GitHub repository tokencanopy/e2a (184 stars, last pushed yesterday), licensed Apache-2.0. It adds 6,823 tokens to every session, about $0.0341 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.