# Send Roadrunner to your agent
Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.
## Fast path
- Download the package from Yavira.
- Extract it into a folder your agent can access.
- Paste one of the prompts below and point your agent at the extracted folder.
## Suggested prompts
### New install

```text
I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete.
```
### Upgrade existing

```text
I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run.
```
## Machine-readable fields
```json
{
  "schemaVersion": "1.0",
  "item": {
    "slug": "roadrunner",
    "name": "Roadrunner",
    "source": "tencent",
    "type": "skill",
    "category": "开发工具",
    "sourceUrl": "https://clawhub.ai/johntheyoung/roadrunner",
    "canonicalUrl": "https://clawhub.ai/johntheyoung/roadrunner",
    "targetPlatform": "OpenClaw"
  },
  "install": {
    "downloadUrl": "/downloads/roadrunner",
    "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=roadrunner",
    "sourcePlatform": "tencent",
    "targetPlatform": "OpenClaw",
    "packageFormat": "ZIP package",
    "primaryDoc": "SKILL.md",
    "includedAssets": [
      "SKILL.md"
    ],
    "downloadMode": "redirect",
    "sourceHealth": {
      "source": "tencent",
      "status": "healthy",
      "reason": "direct_download_ok",
      "recommendedAction": "download",
      "checkedAt": "2026-05-07T17:22:31.273Z",
      "expiresAt": "2026-05-14T17:22:31.273Z",
      "httpStatus": 200,
      "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=afrexai-annual-report",
      "contentType": "application/zip",
      "probeMethod": "head",
      "details": {
        "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=afrexai-annual-report",
        "contentDisposition": "attachment; filename=\"afrexai-annual-report-1.0.0.zip\"",
        "redirectLocation": null,
        "bodySnippet": null
      },
      "scope": "source",
      "summary": "Source download looks usable.",
      "detail": "Yavira can redirect you to the upstream package for this source.",
      "primaryActionLabel": "Download for OpenClaw",
      "primaryActionHref": "/downloads/roadrunner"
    },
    "validation": {
      "installChecklist": [
        "Use the Yavira download entry.",
        "Review SKILL.md after the package is downloaded.",
        "Confirm the extracted package contains the expected setup assets."
      ],
      "postInstallChecks": [
        "Confirm the extracted package includes the expected docs or setup files.",
        "Validate the skill or prompts are available in your target agent workspace.",
        "Capture any manual follow-up steps the agent could not complete."
      ]
    }
  },
  "links": {
    "detailUrl": "https://openagent3.xyz/skills/roadrunner",
    "downloadUrl": "https://openagent3.xyz/downloads/roadrunner",
    "agentUrl": "https://openagent3.xyz/skills/roadrunner/agent",
    "manifestUrl": "https://openagent3.xyz/skills/roadrunner/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/roadrunner/agent.md"
  }
}
```
## Documentation

### roadrunner (rr)

Use rr when the user explicitly wants to operate Beeper Desktop via the local API (send, search, list chats/messages, reminders, focus).
Prefer --agent for agent use (forces JSON, envelope, no-input, readonly).

Safety

Default to read-only commands unless the user explicitly requests a mutation in this turn.
Require explicit recipient (chat ID) and message text before sending.
Confirm or ask a clarifying question if the chat ID is ambiguous.
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.
Use --agent for safe agent defaults: rr --agent --enable-commands=chats,messages,status chats list
Use --readonly to block writes: rr --readonly chats list --json
Use --enable-commands to allowlist: rr --enable-commands=chats,messages chats list --json
Use --envelope for structured errors: rr --json --envelope chats get "!chatid"
Envelope errors may include error.hint with next-step guidance for safe retries.
Never request, paste, or store raw auth tokens in chat. If auth is missing, ask the user to configure it locally.
If sending message text through a shell, avoid interpolation/expansion (e.g. $100/month or !). Prefer --stdin <<'EOF' ... EOF for safe literals.

Setup (once)

rr auth set --stdin (recommended; token saved to ~/.config/beeper/config.json)
rr auth status --check
rr doctor

Common commands

List accounts: rr accounts list --json
Capabilities: rr capabilities --json
Describe command/flags: rr describe messages send --json
Connect metadata: rr connect info --json
Live websocket events (experimental): rr events tail --all --stop-after 30s --json
List contacts: rr contacts list "<account-id>" --json
Search contacts: rr contacts search "<account-id>" "Alice" --json
Search contacts (flag): rr contacts search "Alice" --account-id="<account-id>" --json
Resolve contact: rr contacts resolve "<account-id>" "Alice" --json
Resolve contact (flag): rr contacts resolve "Alice" --account-id="<account-id>" --json
List chats: rr chats list --json
List chats (JSON Lines): rr chats list --jsonl
Search chats: rr chats search "John" --json
Search chats (filters): rr chats search --inbox=primary --unread-only --json
Search chats (activity): rr chats search --last-activity-after="2024-07-01T00:00:00Z" --json
Search by participant name: rr chats search "Jamie" --scope=participants --json
Resolve chat: rr chats resolve "Jamie" --json
Get chat: rr chats get "!chatid:beeper.com" --json
Get chat (bounded participants): rr chats get "!chatid:beeper.com" --max-participant-count=50 --json
Start/resolve DM from merged contact hints: rr chats start "<account-id>" --email "alice@example.com" --full-name "Alice" --json
Default account for commands: rr --account="imessage:+123" chats list --json
List messages: rr messages list "!chatid:beeper.com" --json
List messages (all pages): rr messages list "!chatid:beeper.com" --all --max-items=1000 --json
List messages (download media): rr messages list "!chatid:beeper.com" --download-media --download-dir ./media --json
Search messages: rr messages search "dinner" --json
Search messages (JSON Lines): rr messages search "dinner" --jsonl
Search messages (all pages): rr messages search "dinner" --all --max-items=1000 --json
Search messages (filters): rr messages search --sender=me --date-after="2024-07-01T00:00:00Z" --media-types=image --json
Add/remove reaction: rr messages react "!chatid:beeper.com" "<message-id>" "👍" --json / rr messages unreact "!chatid:beeper.com" "<message-id>" "👍" --json
Tail messages (polling): rr messages tail "!chatid:beeper.com" --interval 2s --stop-after 30s --json
Wait for message: rr messages wait --chat-id="!chatid:beeper.com" --contains "deploy" --wait-timeout 2m --json
Message context: rr messages context "!chatid:beeper.com" "<sortKey>" --before 5 --after 2 --json
Draft message (pre-fill without sending): rr focus --chat-id="!chatid:beeper.com" --draft-text="Hello!"
Draft message from file: rr focus --chat-id="!chatid:beeper.com" --draft-text-file ./draft.txt
Draft with attachment: rr focus --chat-id="!chatid:beeper.com" --draft-attachment="/path/to/file.jpg"
Download attachment: rr assets download "mxc://example.org/abc123" --dest "./attachment.jpg"
Stream attachment bytes: rr assets serve "mxc://example.org/abc123" --dest "./attachment.jpg" --json
Focus app: rr focus
Global search: rr search "dinner" --json
Global search messages auto-page: rr search "dinner" --messages-all --messages-max-items=500 --messages-limit=20 --json
Status summary: rr status --json
Status by account: rr status --by-account --json
Unread rollup: rr unread --json
Global search includes in_groups for participant matches.

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"
Message react/unreact: rr messages react "!chatid:beeper.com" "<message-id>" "👍" / rr messages unreact "!chatid:beeper.com" "<message-id>" "👍"
Upload + send file: rr messages send-file "!chatid:beeper.com" ./photo.jpg "See attached"
Create chat: rr chats create "<account-id>" --participant "<user-id>"
Start chat from merged contact hints: rr chats start "<account-id>" --email "alice@example.com" --full-name "Alice"
Archive/unarchive: rr chats archive "!chatid:beeper.com" / rr chats archive "!chatid:beeper.com" --unarchive
Reminder mutations: rr reminders set "!chatid:beeper.com" "2h" / rr reminders clear "!chatid:beeper.com"
Asset uploads: rr assets upload ./photo.jpg / rr assets upload-base64 --content-file ./photo.b64
For retries on non-idempotent writes, use --request-id and prefer --dedupe-window.
Use --dry-run to validate mutating requests without API write side effects.

Pagination

Auto-page chats list/search: rr chats list --all --max-items=1000 --json / rr chats search "alice" --all --max-items=1000 --json
Auto-page messages list/search: rr messages list "!chatid:beeper.com" --all --max-items=1000 --json / rr messages search "deploy" --all --max-items=1000 --json
Chats: rr chats list --cursor="<oldestCursor>" --direction=before --json
Messages list: rr messages list "!chatid:beeper.com" --cursor="<sortKey>" --direction=before --json
Messages search (max 20): rr messages search "project" --limit=20 --json
Messages search page: rr messages search "project" --cursor="<cursor>" --direction=before --json
Global search message paging (max 20): rr search "dinner" --messages-limit=20 --json
Global search message page: rr search "dinner" --messages-cursor="<cursor>" --messages-direction=before --json

Notes

Requires Beeper Desktop running; token from app settings.
Token is stored in ~/.config/beeper/config.json via rr auth set (recommended). BEEPER_TOKEN overrides the config file.
BEEPER_ACCOUNT sets the default account ID (aliases supported).
rr auth status --check prefers OAuth introspection (/oauth/introspect) when available and falls back to account-list validation on older builds.
Message search is literal word match (not semantic).
rr contacts resolve is strict and fails on ambiguous names; resolve by ID after contacts search when needed.
If a DM title shows your own Matrix ID, use --scope=participants to find by name.
JSON output includes display_name for single chats (derived from participants).
Message JSON includes message_type, linked_message_id, is_sender, is_unread, attachments, and reactions.
downloaded_attachments is only populated when --download-media is used.
rr messages send returns pending_message_id (temporary ID).
Account network may be missing in newer API builds; rr falls back to "unknown" in summaries/search output.
rr assets serve writes raw bytes to stdout unless --dest is provided.
--chat does exact matching and fails on ambiguous matches.
Attachment overrides require --attachment-upload-id; set --attachment-width and --attachment-height together.
--all has a safety cap (default 500 items, max 5000); use --max-items to tune it.
Prefer --json or --jsonl (and --no-input) for automation.
--jsonl emits one JSON object per line and is supported on high-volume list/search commands.
--dry-run/BEEPER_DRY_RUN validates mutating command inputs and prints preview output without sending write API requests.
BEEPER_URL overrides API base URL; BEEPER_TIMEOUT sets timeout in seconds.
JSON/Plain output goes to stdout; errors/hints go to stderr.
Destructive commands prompt unless --force; --no-input/BEEPER_NO_INPUT fails without --force.
Use --fail-if-empty on list/search commands to exit with code 1 if no results.
Use --fields with --plain to select columns (comma-separated).
In bash/zsh, ! triggers history expansion. Prefer single quotes, or disable history expansion (set +H in bash, setopt NO_HIST_EXPAND in zsh).
rr version --json returns features array for capability discovery.
rr capabilities --json returns full CLI capability metadata.
rr events tail depends on experimental /v1/ws support in Beeper Desktop; fall back to rr messages tail when unavailable.
Envelope error codes: AUTH_ERROR, NOT_FOUND, VALIDATION_ERROR, CONNECTION_ERROR, INTERNAL_ERROR.
Retry policy: retry CONNECTION_ERROR with backoff; do not blind-retry AUTH_ERROR/VALIDATION_ERROR; refresh IDs before retrying NOT_FOUND.
Non-idempotent writes: messages send, messages send-file, chats create, chats start, assets upload, assets upload-base64.
Use --request-id/BEEPER_REQUEST_ID to tag envelope metadata for cross-retry attempt tracing.
Use --dedupe-window/BEEPER_DEDUPE_WINDOW to block duplicate non-idempotent writes with repeated request IDs.
Local smoke check: make test-agent-smoke.
## Trust
- Source: tencent
- Verification: Indexed source record
- Publisher: johntheyoung
- Version: 0.17.0
## Source health
- Status: healthy
- Source download looks usable.
- Yavira can redirect you to the upstream package for this source.
- Health scope: source
- Reason: direct_download_ok
- Checked at: 2026-05-07T17:22:31.273Z
- Expires at: 2026-05-14T17:22:31.273Z
- Recommended action: Download for OpenClaw
## Links
- [Detail page](https://openagent3.xyz/skills/roadrunner)
- [Send to Agent page](https://openagent3.xyz/skills/roadrunner/agent)
- [JSON manifest](https://openagent3.xyz/skills/roadrunner/agent.json)
- [Markdown brief](https://openagent3.xyz/skills/roadrunner/agent.md)
- [Download page](https://openagent3.xyz/downloads/roadrunner)