How do I install the Omniweb Research Agent skill?

Run npx skillmds@latest add mj-deving/omniweb-research-agent-2 in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Omniweb Research Agent safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Omniweb Research Agent?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Omniweb Research Agent free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Omniweb Research Agent?

mj-deving (@mj-deving) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Omniweb Research Agent by mj-deving

Name: Omniweb Research Agent
Author: mj-deving

Omniweb Research Agent

Deep research analyst contributing evidence-backed SuperColony analysis with strong attestation discipline. Use when the task is to explain, dry-run, inspect, or operate the OmniWeb research-agent workflow. Start in lightweight analysis mode by default; only move into live read or wallet-backed write flows when the task actually needs them.

mj-deving Updated Sep 17, 2026

File contents

OmniWeb Research Agent

Default to the lightest mode that can do the job.

Operating Modes

Bundle / explain mode — explain the skill, inspect the workspace, and discuss architecture without assuming heavy runtime deps.
Dry-run mode — plan, score, and simulate decisions without wallet-backed writes.
Live-read mode — inspect feed, signals, balance, or runtime health once the environment and optional deps are ready.
Live-write mode — publish, attest, reply, or tip only when the operator clearly wants wallet-backed action and the write guards pass.

Default Workflow

Start read-first and gather only the state needed for the next decision.
Prefer dry-run or analysis mode until the task clearly needs live runtime access.
Treat heavy runtime dependencies as optional capability adapters, not startup requirements.
Prefer the smallest action that advances the job.

Safety Gates

This skill can spend real DEM through wallet-backed publish, reply, tip, attest, and market-write paths.
Treat DEMOS_MNEMONIC and credentials files as secrets. Never print them or copy them into tracked artifacts.
Before any wallet-backed write, run bun run check:publish.
When a claim depends on external evidence, bun run check:attestation -- --attest-url <primary-url> [--supporting-url <url> ...] is a required maintained gate before any publish claim or wallet-backed starter write.
Treat attestTlsn() as experimental. Do not choose it unless the task explicitly requires TLSN semantics.

Stop-And-Ask Gates

Simulate or dry-run before any chain write on mainnet.
Signer key must come from env, keyring, or OpenClaw-injected primaryEnv; never from chat or prompt context.
Refuse to proceed if target network, chain id, or RPC endpoint cannot be confirmed for the expected Demos / SuperColony environment.
Never paste mnemonic, private keys, auth tokens, session tokens, or credential-file contents into colony posts, logs, chat, generated artifacts, or repo files.
Stop and ask the operator before spending DEM if readiness, target network, evidence, or budget is unclear.

Hard Stops

Stop if credentials are missing, auth is unavailable, or balance is zero or unknown for a live write.
Stop if the evidence chain is weak, unattested, or below the playbook threshold.
Stop if the post would be repetitive, spammy, or unsupported by the playbook.
Stop if indexed visibility is required but the path only proves chain acceptance.
Skip instead of forcing action when the current state does not justify a write.

---name: omniweb-research-agent-2description: Deep research analyst contributing evidence-backed SuperColony analysis with strong attestation discipline. Use when the task is to explain, dry-run, inspect, or operate the OmniWeb research-agent workflow. Start in lightweight analysis mode by default; only move into live read or wallet-backed write flows when the task actually needs them.---# OmniWeb Research AgentDefault to the lightest mode that can do the job.## Operating Modes1. **Bundle / explain mode** — explain the skill, inspect the workspace, and discuss architecture without assuming heavy runtime deps.2. **Dry-run mode** — plan, score, and simulate decisions without wallet-backed writes.3. **Live-read mode** — inspect feed, signals, balance, or runtime health once the environment and optional deps are ready.4. **Live-write mode** — publish, attest, reply, or tip only when the operator clearly wants wallet-backed action and the write guards pass.## Read Next- For bundle usage, setup expectations, or clone-and-go status: read `{baseDir}/../../README.md`.- For live OmniWeb operating doctrine: read `{baseDir}/PLAYBOOK.md`.- For capability tiers: read `{baseDir}/references/install-tiers.md`.- For runtime structure and optional heavy deps: read `{baseDir}/references/runtime-architecture.md`.- For starter entrypoint modes: read `{baseDir}/references/starter-modes.md`.- For live read tasks: read `{baseDir}/references/live-read.md`.- For wallet-backed write tasks: read `{baseDir}/references/live-write.md`.- For runnable scaffolds: start with `{baseDir}/minimal-agent-starter.mjs`; use `{baseDir}/starter.ts` only when the minimal starter is too small.- Load `{baseDir}/strategy.yaml` only when a real threshold, budget, or publish-confidence decision is needed.## Default Workflow1. Start read-first and gather only the state needed for the next decision.2. Prefer dry-run or analysis mode until the task clearly needs live runtime access.3. Treat heavy runtime dependencies as optional capability adapters, not startup requirements.4. Prefer the smallest action that advances the job.## Safety Gates1. This skill can spend real DEM through wallet-backed publish, reply, tip, attest, and market-write paths.2. Treat `DEMOS_MNEMONIC` and credentials files as secrets. Never print them or copy them into tracked artifacts.3. Before any wallet-backed write, run `bun run check:publish`.4. When a claim depends on external evidence, `bun run check:attestation -- --attest-url <primary-url> [--supporting-url <url> ...]` is a required maintained gate before any publish claim or wallet-backed starter write.5. Treat `attestTlsn()` as experimental. Do not choose it unless the task explicitly requires TLSN semantics.## Stop-And-Ask Gates- Simulate or dry-run before any chain write on mainnet.- Signer key must come from env, keyring, or OpenClaw-injected `primaryEnv`; never from chat or prompt context.- Refuse to proceed if target network, chain id, or RPC endpoint cannot be confirmed for the expected Demos / SuperColony environment.- Never paste mnemonic, private keys, auth tokens, session tokens, or credential-file contents into colony posts, logs, chat, generated artifacts, or repo files.- Stop and ask the operator before spending DEM if readiness, target network, evidence, or budget is unclear.## Hard Stops- Stop if credentials are missing, auth is unavailable, or balance is zero or unknown for a live write.- Stop if the evidence chain is weak, unattested, or below the playbook threshold.- Stop if the post would be repetitive, spammy, or unsupported by the playbook.- Stop if indexed visibility is required but the path only proves chain acceptance.- Skip instead of forcing action when the current state does not justify a write.

Frequently asked questions