relayBLOCK
Infrastructure for coding agents
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
Tired of copy/pasting from Claude Code into Slack so your coworker can paste it into their agent?
Ever put an important rule in a skill or AGENTS.md, only for the agent to completely ignore it?
You and your teammates agents ever build the same thing? Wondered whether your coworker's agent made that change intentionally? How many times have you typed "babysit this PR until CI passes"?
Agent Relay is an open-source toolkit for problems like these. It gives engineering teams infrastructure for running coding agents together without replacing the agents and tools they already use.
Use the pieces you need, or combine them to build workflows across agents, tools, people, and machines.
Messaging
Claude/Codex/etc can talk directly through shared channels, threads, DMs, files, search, and real-time events. Agents can run on differe
1e3f291f57baOBSERVED · 2026-10-12Install
Commands as the repository documents them. They are shown, not run.
npm install -g agent-relay
npm install -g agent-relay
npm install -g mycli
git clone https://github.com/badlogic/agent-tools.git
npm install -g agent-relay
npm install -g agent-relay
Host compatibility
What the documentation claims. We have not run a compatibility test.
| Host | Status | Notes |
|---|---|---|
| claude-code | mentioned | |
| codex | mentioned | |
| copilot | mentioned | |
| cursor | mentioned | |
| gemini-cli | mentioned | |
| openclaw | mentioned | |
| windsurf | mentioned |
What it tells the agent
The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.
# Writing Relayflows
Use when authoring a Relayflows flow (@relayflows/surface / @relayflows/sdk, the journal-based v2 engine — the CLI is `flows`, package versions 2.0.x) in TypeScript or YAML/JSON. Covers the three-rung ladder (run/llm/agent), the resident verbs (human/dispatch/done), verification gates, TypeScript vs YAML authoring, per-step cli/model selection and its resolution order, flows.json, and `flows check`/`run`/`resume` with their real refusal shapes and exit codes. Not for the older, unrelated `@relayflows/core` WorkflowBuilder engine (`.pattern('dag')`/.agent()/.step() chains) that `writing-agent-relay-workflows` and `migrating-persona-to-relayflow` cover — that's a different product despite the similar name.
## Overview
Relayflows turns a coding-agent task into steps a journal can inspect, verify, and resume. A flow is data (YAML/JSON) or code (TypeScript) that compiles to the same journal-backed kernel spec. Every effect is journaled before it's treated as real — a journal write that fails fails the step, with no silent fallback.
**Name collision warning.** This repo also has skills for an older, unrelated engine that is _also_ casually called "Relayflow" (singular) — `@relayflows/core`'s `WorkflowBuilder`, a chained builder (`workflow('name').pattern('dag').agent(...).step(...).run()`). That's `writing-agent-relay-workflows` and `migrating-persona-to-relayflow`'s territory. This skill is the **v2** engine: `@relayflows/surface`'s `flow()` function and the YAML/JSON dialect compiled by `@relayflows/sdk`. If you see `.pattern(`, `.agent(` as a chained builder call, or `ctx.workflow.run()`, you're in the other engine — stop and use one of those skills instead.
## When to use this skill
- Writing a new `.flow.ts` or `.flow.yaml`/`.flow.json` for the `flows` CLI (package `@relayflows/sdk`, binary name `flows`).
- Deciding whether a step needs `run` (shell), `llm` (bare model call), or `agent` (harnessed coding agent in a workspace).
- Wiring up `cli`/`model` for an `agent` or `llm` step, in either language.
- Debugging a `REFUSED [...]` message from `flows check` or `flows run`.
- Choosing between TypeScript and YAML for a given flow.
## The ladder
Three step verbs, one per rung — never more (`packages/sdk/src/spec.ts`, `export type StepType = 'deterministic' | 'llm' | 'agent';`):
1. **`run` / `deterministic`** — a shell command. No model. Implicit gate is `exit_code == 0`.
2. **`llm` / `llm`** — a bare model call. Prompt in, verified output out. No workspace, no tool use.
3. **`agent` / `agent`** — a harnessed coding agent in a workspace. Returns `{ summary, artifacts }`, not raw text.
Plus four resident verbs that aren't ladder rungs: `human` (durable approval), `dispatch` (hand off to a child flow), `done` (typed finish), and in YAML, `on`/triggers (event entry points — out of scope for this skill).
Most flows only need `run` and `llm`. Climb to `agent` once a step needs hands on a real workspace.
## Two ways to author the same thing
### TypeScript
```ts
import { flow } from '@relayflows/surface';
export default flow('hello', async (f) => {
const greeting = await f.run('echo "Hello from Relayflows"');
console.log(greeting.trim());
const answer = await f.agent('greeter', {
task: 'Reply with one short hello sentence. Do not use tools or modify files.',
cli: 'claude',
model: 'claude-sonnet-4-6',
});
console.log(answer.summary);
f.done('success');
});
```
### YAML
```yaml
version: '0.1.0'
name: hello
steps:
- id: greeting
type: deterministic
command: 'echo "Hello from Relayflows"'
- id: greeter
type: agent
dependsOn: [greeting]
instruction: 'Reply with one short hello sentence. Do not use tools or modify files.'
cli: claude
model: claude-sonnet-4-6
```
## The real `Ctx` contract (TypeScript)
### `packages/surface/src/context.ts`, current as of `origin/main@86a2ec2`:
```ts
export interface AgentResult {
summary: string;
artifacts: string[];
}
export interface AgentOptions {
task: string;
workspace?: string;
cli?: string;
model?: string;
}
export interface Ctx {
run(command: string): Step<string>;
llm(strings: TemplateStringsArray, ...values: unknown[]): Step<string>;
llm(
prompt: string,
options: { output: Record<string, unknown>; cli?: string; model?: string }
): Step<unknown>;
agent(name: string, options: AgentOptions): Step<AgentResult>;
human(question: string, options: { to: string }): Promise<boolean>;
dispatch<T>(flow: string, input: unknown): Promise<T>;
done(reason: RunCompletionReason): void;
cloud: CloudHelper;
slack: SlackHelper;
}
```
## The real step shapes (YAML/JSON, `packages/sdk/src/spec.ts`)
### ```ts
```ts
interface DeterministicStepSpec {
type: 'deterministic';
id: string;
command: string;
dependsOn?: string[];
timeoutMs?: number;
verification?: VerificationSpec; // omit for implicit exit_code
}
interface LlmStepSpec {
type: 'llm';
id: string;
prompt: string;
dependsOn?: string[];
verification?: OutputVerificationSpec;
model?: string;
cli?: string;
}
interface AgentStepSpec {
type: 'agent';
id: string;
instruction: string;
dependsOn?: string[];
verification?: OutputVerificationSpec;
agent?: string; // selects a named FlowSpec.agents entry
cli?: string;
model?: string;
surfaces?: { workspace?: { surface: string }[]; streams?: { stream: string }[]; external?: string[] };
recoveryMode?: 'reset' | 'inspect' | 'manual'; // default 'reset'
permissions?: {
fileGlobs?: string[];
networkAllowlist?: string[];
accessPreset?: 'readonly' | 'readwrite';
};
}
interface FlowSpec {
version: string; // required, e.g. '0.1.0' — not optional
name?: string;
cli?: string; // flow-level CLI default
agents?: Record<string, { cli: string; model: string }>; // both fields required
steps: StepSpec[];
budget?: { maxTokensIn?: number; maxTokensOut?: number; maxDollars?: string };
}
```
#Trust audit
BLOCKgrade F · trust 26/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | WARN |
| L1 | Static analysis of the code | FAIL |
| L2 | Instruction surface (what it tells the agent) | FAIL |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (8 observation(s))
- Network
- declared (13 observation(s))
- Shell
- declared (2 observation(s))
- Dependencies
- not all pinned
- Secrets in source
- found
Findings (25)
| Told the user to open an observer URL built from the workspace key | That is an admin credential in a query string, and the r
curl -sS -X POST "https://agentrelay.com/cloud/api/v1/workspaces/<id>/integrations/connect-session" \
let settings = br#"{"permissions":{"deny":["Exec(sudo)"]}}"#;.agents/skills/adding-swarm-patterns
.claude/skills/adding-swarm-patterns
.claude/skills/emil-design-eng
.claude/skills/review-animations
callback_url: workers
callback_url: self
callback_url: "http://127.0.0.1:3889/api/agent-result".to_string(),
callback_url: "http://127.0.0.1:3889/api/agent-result".to_string(),
pub callback_url: String,
"endpoint": "http://127.0.0.1:4096",
"endpoint": "http://127.0.0.1:4096",
resolve_relaycast_base_url(Some("http://127.0.0.1:8787")),"http://127.0.0.1:8787"
.unwrap_or("http://127.0.0.1:3889/api/agent-result")for (const token of ['sk-proj-0123456789abcdefghijklmnop', 'sk-ant-0123456789abcdefghijklmnop']) {let secret = "raw-ownership-proof-must-not-appear";
let token = "at_live_metadata_fixture";
token: "fixture-task-callback".into(),
api_key: 'sk_live_should_not_leak_1234567890',
api_key: 'sk_live_should_not_leak_1234567890',
const githubToken = 'github_pat_0123456789abcdef0123456789abcdef';
const token = 'github_pat_abcdefghijklmnopqrstuvwxyz123456';
Gates applied: critical_finding, no_behavioural_pass, undeclared_transfer.
1e3f291f57bafull audit observations/trust-audit/skill/agentworkforce__relay.json · Report an issue / request a re-scanAudit history
Every audit this skill has had.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-12 | 1e3f291f57ba | BLOCK | F | 26 | first audit |
Questions
What does the relay skill do?
Infrastructure for coding agents
Is relay safe to install?
No — not without reading the findings first. The audit graded it F (26/100) and found 3 critical or high issues in the source. Each one is listed on this page with the file and line it is on.
What can relay access on my machine?
The audit observed that it reaches the network, runs shell commands and reads or writes files. Each of those is consistent with what it says it does. Secrets in the source: found — see the findings.
What do I need installed to use relay?
Its own instructions reference agent, llm and run. Dependencies are not all pinned to exact versions, so what installs today may differ tomorrow.
Which assistants does relay work with?
Its documentation mentions claude-code, codex, copilot, cursor, gemini-cli, openclaw and windsurf. That is what the text claims, not a compatibility test we ran.
How current is this page?
The grade is for one exact copy of the source (1e3f291f57ba), read on 2026-10-12. The repository is watched, and a new audit runs when it changes — this is the first audit.