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/clawrun-sh/clawrun/agents-mdgit clone --depth 1 https://github.com/clawrun-sh/clawrunWhat 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 | $0.01508 | $0.01508 |
| Opus 5 | $0.00754 | $0.00754 |
| Sonnet 5 | $0.00302 | $0.00302 |
| Haiku 4.5 | $0.00151 | $0.00151 |
Grade A, and why
clawrun 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 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.
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 — 168 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — Agent Interface & Implementation Guide
Agent Interface
Defined in packages/agent/src/types.ts. Every agent implementation must implement this interface.
interface Agent {
readonly id: string;
readonly name: string;
readonly channelsConfigKey: string;
readonly daemonPort: number;
// Provisioning
provision(sandbox, root, opts: ProvisionOpts): Promise<void>;
// Messaging
sendMessage(sandbox, root, message, opts?): Promise<AgentResponse>;
streamMessage(sandbox, root, message, writer: UIMessageStreamWriter, opts?): Promise<void>;
// Conversation threads
listThreads(sandbox, root, opts?): Promise<ThreadInfo[]>;
getThread(sandbox, root, threadId, opts?): Promise<UIMessage[]>;
// Daemon and cron
getDaemonCommand(root, opts?): DaemonCommand;
listCronJobs(sandbox, root, opts?): Promise<CronJob[]>;
// Tools
getAvailableTools(): Tool[];
getEnabledTools(agentDir): Tool[];
getToolDomains(agentDir): Tool[];
// Configuration and setup
getMonitorConfig(root): MonitorConfig;
getProviders(): ProviderInfo[];
getDefaultModel(provider): string;
getCuratedModels(provider): CuratedModel[];
getModelsFetchEndpoint(provider, apiUrl?): { url, authHeader } | null;
getSupportedChannels(): ChannelInfo[];
writeSetupConfig(agentDir, data: AgentSetupData): void;
readSetup(agentDir): { provider?; channels?; cost? } | null;
// Bundling and dependencies
getLocalOwnedFiles(): string[];
getBundleFiles(): string[];
getBinaryBundlePaths(): string[];
getInstallDependencies(): Record<string, string>;
getSeedDirectory(): string | null;
// Dashboard API (called via daemon HTTP endpoints)
getStatus(sandbox, root, opts?): Promise<AgentStatus>;
getConfig(sandbox, root, opts?): Promise<AgentConfig>;
listTools(sandbox, root, opts?): Promise<ToolsResult>;
createCronJob(sandbox, root, job, opts?): Promise<CronJob>;
deleteCronJob(sandbox, root, id, opts?): Promise<void>;
listMemories(sandbox, root, query?, opts?): Promise<MemoryEntryInfo[]>;
createMemory(sandbox, root, entry, opts?): Promise<void>;
deleteMemory(sandbox, root, key, opts?): Promise<void>;
getCostInfo(sandbox, root, opts?): Promise<CostInfo>;
runDiagnostics(sandbox, root, opts?): Promise<DiagResult[]>;
}
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 · 168 lines · 1,508 tokens per session scan A 65b4644c44a8
clawrun AGENTS.md is an instructions file published in the GitHub repository clawrun-sh/clawrun (94 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 1,508 tokens to every session, about $0.0075 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.
Other instructions, from other repositories
zeroclaw AGENTS.md
AGENTS.md instructions for zeroclaw-labs/zeroclaw, covering agents.md - zeroclaw, single source of truth, safety and privacy, working rules and user-facing text.
zeroclaw CLAUDE.md
Claude Code instructions for zeroclaw-labs/zeroclaw, covering claude.md — zeroclaw (claude code), claude code settings, hooks and slash commands.
agentos CLAUDE.md
Claude Code instructions for rivet-dev/agentos, covering agentos, boundaries, security model, sqlite schema ownership and runtime and registry.
agent-sandbox AGENTS.md
Instructions for kubernetes-sigs/agent-sandbox, covering agents.md, project summary, repository layout, agent skills and build, test, lint.
agent-sandbox copilot-instructions.md
Instructions for kubernetes-sigs/agent-sandbox: Project Context & Architecture: Refer to AGENTS.md for full project background, module layout, toolchain versions, and core conventions.
SmolVM AGENTS.md
Instructions for CelestoAI/SmolVM, covering smolvm context, 🚀 project overview, 🧪 development, key commands and release checklist.