Rightbrain Agent Skill

An Agent Skill for safely building and operating Rightbrain Tasks, TaskAgents, knowledge, tools, and automations.

The skill reads Rightbrain's live documentation and OpenAPI contract at execution time. It does not bundle an API reference or require API-spec syncs.

How it works

• Rightbrain CLI owns interactive login and session refresh.
• rb-api.mjs gives agents one safe interface for every API operation, including scoping, pagination, streaming, uploads, dry runs, and secret redaction.
• Live docs provide the current endpoint contract: llms.txt, Docs MCP, and OpenAPI.

Quick start

1. Install the skill.
2. Install and authenticate the CLI:

npm install --global rightbrain@latest
rightbrain login --non-interactive

3. Ask your agent to use the Rightbrain skill.

The skill verifies the selected organization, project, and production API host before making changes. Headless environments may use secret-managed API keys or OAuth client credentials instead of an interactive session.

Safety

Writes show their target and effect first. Deletion, public sharing, credentials, permissions, automation, external writes, and potentially expensive runs require explicit confirmation.

Never paste secrets into prompts, command arguments, issues, or logs. Structured responses redact known secret fields; arbitrary model and tool output cannot be guaranteed secret-free. See SECURITY.md.

Development

node --test test/*.test.mjs

See CONTRIBUTING.md, CHANGELOG.md, and LICENSE.