alvis-agent-builderBuild and update OpenClaw agents by designing personas, rules, autonomy, and generating complete agent workspace files with safety and behavior checks.
Install via ClawdBot CLI:
clawdbot install alvisdunlop/alvis-agent-builderGrade Fair — based on market validation, documentation quality, package completeness, maintenance status, and authenticity signals.
Sends data to undocumented external endpoint (potential exfiltration)
POST → https://api.heybossai.com/v1/pilot`Potentially destructive shell commands in tool definitions
eval (Calls external URL not in known-safe list
https://api.heybossai.com/v1/pilot`AI Analysis
The skill definition itself contains no malicious instructions, but the rule-based signals indicate it triggers external API calls to an undocumented endpoint (api.heybossai.com) and potentially unsafe shell commands. While the skill's purpose is agent design, the execution environment appears to have hidden behaviors not described in the provided definition, posing a risk of data exfiltration or system compromise.
Generated Oct 6, 2026
An individual power user designs a personal OpenClaw agent to triage email, draft replies, and manage a daily heartbeat briefing across Telegram DM. The builder interview sets autonomy to Operator so the agent drafts but always asks before sending outbound messages. Acceptance tests validate ask-before-send and secret-redaction behavior.
A SaaS startup builds an OpenClaw agent that monitors a support Discord and answers common questions using tool-first verification. The team configures group chat etiquette and hard prohibitions against promising refunds or feature dates. Guardrails ensure the agent escalates destructive or external actions to a human.
A digital agency productizes agent builds for clients, creating branded personas that operate on WhatsApp and iMessage with warm, concise tone rules. Each client gets a tailored SOUL.md and AGENTS.md with autonomy tuned to their risk tolerance. The agency iterates on failure modes like overreach and verbosity after launch.
An e-commerce ops team deploys an Autopilot-level agent that watches store metrics and posts heartbeat summaries to a group channel. MEMORY.md tracks recurring incident categories while loop-breaker rules prevent runaway retries. Hard prohibitions block the agent from touching production configs without approval.
A fintech or healthcare team builds an Advisor-mode agent that only suggests actions and never executes transactions or sends outbound messages. The workspace enforces strict verification requirements, secret redaction, and stop-on-CLI-error rules for auditability. Acceptance tests confirm the agent refuses overreach under pressure.
A service business that interviews clients, designs custom OpenClaw personas and operating rules, and delivers complete workspace files. They charge project fees plus retainers for iterating on failure modes, guardrails, and heartbeat plans. Margins come from templatized delivery combined with bespoke tuning.
A product business selling pre-built OpenClaw agent blueprints (support bot, ops monitor, sales concierge) with SOUL.md, AGENTS.md, and test suites included. Buyers customize templates instead of starting from scratch, accelerating time-to-value. Upsells include premium guardrail packs and industry-specific rosters.
A subscription service that hosts, monitors, and continuously tunes OpenClaw agents for small businesses. The provider runs heartbeat reviews, updates guardrails, and reports on autonomy incidents. Recurring value comes from ongoing iteration rather than one-time builds.
💬 Integration Tip
Always read references/openclaw-workspace.md and references/templates.md before generating files so workspace layout and heartbeat rules stay canonical. Run the Phase 4 acceptance tests immediately after generation to catch guardrail gaps before deployment.
Scored Aug 30, 2026
Audited Apr 17, 2026 · audit v1.0
Meta-skill for AI agent self-improvement. Analyzes runtime logs to detect error patterns, regressions, and inefficiencies, then generates structured improvem...
Stop waiting for prompts. Keep working.
Turn OpenClaw into a learning-loop agent with seeded workspace rules, skill promotion, reflective memory, and proactive maintenance.
Complete toolkit for creating autonomous AI agents and managing Discord channels for OpenClaw. Use when setting up multi-agent systems, creating new agents, or managing Discord channel organization.
Billions decentralized identity for agents. Link agents to human identities using Billions ERC-8004 and Attestation Registries. Verify and generate authentic...
AI agent for finance search agent tasks