Borrowing it
Nothing to install: this file belongs to nkiyohara/corresync. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/nkiyohara/corresync/main/AGENTS.mdgit clone --depth 1 https://github.com/nkiyohara/corresyncWrote 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/nkiyohara/corresync/agents-md)<a href="https://agentmods.dev/instructions/nkiyohara/corresync/agents-md"><img src="https://agentmods.dev/badge/instructions/nkiyohara/corresync/agents-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/nkiyohara/corresync/agents-md"><img src="https://agentmods.dev/badge/instructions/nkiyohara/corresync/agents-md.svg" alt="Reviewed on agentmods" width="80" 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 | $0.01238 | $0.01238 |
| Opus 5.5 | $0.00495 | $0.00495 |
| Sonnet 5.5 | $0.00248 | $0.00248 |
| Haiku 4.5 | $0.00124 | $0.00124 |
Grade A, and why
corresync 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 10d 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 — 106 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Repository instructions
Scope
Build Corresync, a local-first, provider-neutral mail, calendar, task, and messaging CLI and MCP server for accounts the signed-in human already controls. Mail and calendar are in scope across providers, including a provider-native Teams or Google Meet join link provisioned as a property of one calendar event. Tasks and the evidence-gated messaging domain are defined in ADRs 0028 and 0034. See ADR 0008 for the accepted scope and ADR 0011 for the rename.
Scope is not capability. Development-route claims require synthetic fixture contract tests and remain live-unobserved until a documented opt-in live observation exists. A route that lacks complete release evidence remains unreachable unless the maintainer explicitly enabled its source-owned preview under ADR 0034. Teams Graph/Web and Mattermost are such previews; missing evidence remains missing and must be reported. Slack stays disabled.
Teams calls, recordings, presence surveillance, and meeting lifecycle management stay out of scope. Provider-neutral chat and channel messaging is permitted only through the bounded contracts and immutable release gates in ADR 0034; it does not broaden the calendar operation in ADR 0005. Hosted relays, multi-user servers, unattended credential login, TLS interception, tenant-wide or administrative access, and any bypass of authentication, MFA, Conditional Access, a disabled service, or a permission the user does not already have are permanently out of scope. Microsoft Graph is never an implicit dependency, an automatic fallback, or a capability probe; it is used only on explicit user selection or with an authorization the user already granted.
Architecture invariants
- Dependencies point inward: adapters and transports may depend on application ports; the domain must not import them. Provider adapters translate and hold no policy.
- CLI and MCP must call the same typed application use cases. No provider gets an escape hatch, an arbitrary action, or a generic property API.
- Where a web adapter is used, authentication is interactive and browser-owned. Never request or persist a password, app-specific password, OAuth token, cookie, or refresh token in core, and never introduce TLS interception. The configuration schema cannot represent a secret; a standards provider that needs one reads it through the consent-gated external credential port in ADR 0012.
- Discovery is credential-free, requires valid TLS, and never triggers an administrator-consent or admin-review flow. Automatic selection never starts a Graph or managed Google authorization.
- Accounts are isolated: opaque stable identity independent of address and alias, with separate sessions, profiles, cursors, caches, and audit context. Reads may aggregate; every write resolves one exact account and container first.
- Consequential writes use the server-enforced preview/commit protocol, and the approval token stays bound to the exact previewed account, target, and payload.
- Capabilities are observed per account after sign-in, never inferred from branding. Degradation is reported explicitly and never normalized away.
- Import scanning is read-only and never reuses another application's secrets. Monitoring, agent dispatch, and data egress are separate opt-ins that default to disabled, and MCP can never enable them.
- MCP annotations describe effects but never replace core policy checks.
- Live mailbox tests are opt-in and cannot run in the default test command or CI.
- Fixtures are synthetic and contain no credentials or personal data.
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.
- 10d ago Changed · +3 lines · +33 tokens per session 9f31347b0b06
- 1mo ago Changed · -7 lines · -100 tokens per session a0718b33dac7
- 2mo ago First seen · 110 lines · 1,305 tokens per session scan A d436aab5e2be
corresync AGENTS.md is an instructions file published in the GitHub repository nkiyohara/corresync (2 stars, last pushed 2d ago), licensed Apache-2.0. It adds 1,238 tokens to every session, about $0.0050 per session on Opus 5.5. 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
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, cache eligibility and rollout and applying the notice (cutover).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode copilot-instructions.md
Copilot instructions for microsoft/vscode, covering vs code copilot instructions, project overview, root folders, core architecture (src/ folder) and built-in extensions (extensions/ folder).
deepseek-harness AGENTS.md
AGENTS.md instructions for deepseek-ai/deepseek-harness, covering agents.md, pre-stable apis and released session data, repository layout, commands and host sandbox failures.