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/swayamg20/agentrelay/agents-mdgit clone --depth 1 https://github.com/swayamg20/AgentRelayWrote 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/swayamg20/agentrelay/agents-md)<a href="https://agentmods.dev/instructions/swayamg20/agentrelay/agents-md"><img src="https://agentmods.dev/badge/instructions/swayamg20/agentrelay/agents-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 | $0.00920 | $0.00920 |
| Opus 5 | $0.00460 | $0.00460 |
| Sonnet 5 | $0.00184 | $0.00184 |
| Haiku 4.5 | $0.00092 | $0.00092 |
Grade A, and why
AgentRelay 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 4d 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 — 87 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AgentRelay repository instructions
Prime directive
Understand the system before changing it. Write the smallest clear solution that fully solves the requested problem. Do not trade readability for cleverness, and do not add abstractions, configuration, or defensive machinery without a concrete need.
Before editing
- Inspect
git status, the relevant implementation, its callers, and nearby tests. - Preserve unrelated work, including untracked files and stashes.
- Read the relevant design source:
docs/architecture.mdfor current and target component boundaries.docs/hld.mdfor the implemented mailbox and Relay control-plane flow.docs/lld.mdfor current schemas, routes, tools, and known gaps.docs/rfcs/001-agentrelay-node-and-missions.mdfor the next architecture.
- Trace cross-package behavior through
protocol/,relay/,mcp-server/, andnode/as applicable. A contract is not understood until its producer and consumer have both been checked. - Treat active code and tests as current behavior. Treat an accepted RFC as target behavior. If they disagree, state the discrepancy instead of silently choosing.
Implementation style
- Make the smallest coherent change that satisfies the request.
- Prefer direct code and existing patterns over speculative reuse.
- Extract a helper only when it names a real rule, removes meaningful duplication, or makes a failure boundary easier to see.
- Keep functions focused, names concrete, and data flow easy to follow.
- Comments explain why, constraints, or non-obvious failure modes. Do not narrate what the code already says.
- Do not perform unrelated cleanup or add features "while you are here."
- Do not add a dependency unless the task genuinely needs it. Explain any addition.
- Preserve public HTTP, JSON-RPC, MCP, CLI, database, and config contracts unless the task explicitly changes them.
- Behavioral changes require focused tests for the changed contract. Do not add speculative tests for unrelated or hypothetical behavior.
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.
- 4d ago First seen · 87 lines · 920 tokens per session scan A aa9682a92431
AgentRelay AGENTS.md is an instructions file published in the GitHub repository swayamg20/AgentRelay (5 stars, last pushed 14d ago), licensed MIT. It adds 920 tokens to every session, about $0.0046 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 2026-08-31.
Other instructions, from other repositories
a2amesh AGENTS.md
AGENTS.md instructions for oaslananka/a2amesh, covering a2a mesh agent map, what this repository is, setup, build and test and repository layout.
a2a-demos copilot-instructions.md
Instructions for capiscio/a2a-demos, covering a2a-demos - github copilot instructions, ⛔ absolute rules - no exceptions, 1. all work via pull requests, 2. no watch/blocking commands and 🚨 critical: read first.
matt-skills-with-to-goal CLAUDE.md
Instructions for tt-a1i/matt-skills-with-to-goal: Skills are organized into bucket folders under skills/.
ephemeral-sandbox CLAUDE.md
Instructions for Ephemeral-AI-Lab/ephemeral-sandbox, covering claude.md, project, engineering practice (required), build & test and sandbox tools.
adk CLAUDE.md
Claude Code instructions for inference-gateway/adk, covering claude.md, what this repo is, commands, architecture and server (server/).
coral CLAUDE.md
Instructions for cdknorow/coral, covering claude.md - coral go, mission, testing, go unit tests and legacy parity tools (historical reference).