Roadrunner

PassAudited by VirusTotal on May 12, 2026.

Overview

Type: OpenClaw Skill Name: roadrunner Version: 0.17.0 The roadrunner skill is a CLI tool for interacting with Beeper Desktop, providing capabilities to manage chats, messages, and contacts. The SKILL.md file contains extensive safety instructions for the AI agent, explicitly directing it to use read-only modes by default, avoid leaking raw authentication tokens or large data dumps, and use structured error handling. The tool's capabilities (e.g., reading messages, downloading media to ./media) are entirely consistent with its stated purpose as a messaging client, and no evidence of malicious intent or obfuscation was found.

Findings (0)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

If invoked incorrectly, the agent could send, edit, or react to Beeper messages, but the skill’s instructions require explicit user intent and recipient details.

Why it was flagged

The CLI can perform high-impact chat actions such as sending and editing messages, but the instructions explicitly require user-directed mutations and default read-only use.

Skill content
Default to read-only commands unless the user explicitly requests a mutation in this turn. ... Mutations (explicit user request only) - Message send: `rr messages send "!chatid:beeper.com" "Hello!"` - Message edit: `rr messages edit "!chatid:beeper.com" "<message-id>" "Updated text"`
Recommendation

Use this skill only for specific Beeper tasks, verify chat IDs and message text before sending, and prefer the documented `--agent`, `--readonly`, and command allowlist options.

What this means

The configured CLI can act with the permissions of your Beeper account and access your chats, contacts, and messages.

Why it was flagged

The skill uses a local Beeper authentication token, which is expected for operating Beeper Desktop, and it instructs the agent not to request or expose raw tokens.

Skill content
`rr auth set --stdin` (recommended; token saved to `~/.config/beeper/config.json`) ... Never request, paste, or store raw auth tokens in chat. If auth is missing, ask the user to configure it locally.
Recommendation

Configure authentication locally, do not paste tokens into chat, protect the local config file, and revoke or rotate the token if you no longer trust the environment.

What this means

Installing the skill means trusting the external Roadrunner CLI to handle Beeper data and account actions correctly.

Why it was flagged

The skill depends on installing an external CLI from a Homebrew tap or Go module. That is expected for an rr-based skill, but it is still a trust dependency.

Skill content
brew | formula: johntheyoung/tap/roadrunner | creates binaries: rr ... go | module: github.com/johntheyoung/roadrunner/cmd/rr@v0.17.0 | creates binaries: rr
Recommendation

Install from the documented source, prefer the pinned Go module when appropriate, and review the upstream project before granting it access to your Beeper account.

What this means

Private chat content may be read by the agent for the requested task and could be summarized back to the user.

Why it was flagged

The skill can retrieve private chat lists and message data into the agent context, and the instructions appropriately warn against raw disclosure.

Skill content
Never paste raw rr command output (JSON dumps, chat lists, etc.) into outgoing messages. Treat tool output as private; summarize or extract only what the user needs.
Recommendation

Keep requests narrowly scoped, avoid broad message dumps unless needed, and ask the agent to summarize only the specific information required.