Manus

PassAudited by VirusTotal on May 12, 2026.

Overview

Type: OpenClaw Skill Name: manus Version: 1.3.1 The manus skill is a legitimate integration for the Manus AI platform, providing tools to create, monitor, and download results from autonomous agent tasks. The implementation in `scripts/manus.sh` follows good security practices, such as sanitizing filenames before downloading and using `jq` to safely handle JSON payloads, with no evidence of malicious intent or data exfiltration.

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

A Manus task may run autonomously based on the prompt and could produce actions, results, or provider-side costs the user did not fully anticipate.

Why it was flagged

The skill intentionally delegates work to an external autonomous agent that can use tools; this matches the stated purpose but is still a capability users should notice.

Skill content
Use the Manus API to create autonomous AI tasks. Manus can browse the web, use tools, and deliver complete results
Recommendation

Use clear, bounded prompts, avoid delegating sensitive or irreversible decisions, and review Manus outputs before relying on them.

What this means

Anyone or any process with access to the key could use the Manus account, view tasks available to the key, or create new tasks that may consume credits.

Why it was flagged

The skill requires a Manus API key and uses it to create, list, and read tasks; this is expected for the integration but grants account-level access to Manus operations.

Skill content
Header: `API_KEY: <your-key>` ... Set via: - `MANUS_API_KEY` env var - Or `skills.manus.apiKey`
Recommendation

Store the key securely, use the least-privileged key available, rotate it if exposed, and monitor Manus account usage.

What this means

Information included in prompts or generated outputs may be processed and stored by Manus, and shareable links may broaden access depending on Manus link controls.

Why it was flagged

User prompts are sent to the external Manus service, and the recommended workflow creates shareable task links; this is disclosed but affects data boundaries.

Skill content
"prompt": "Your task description here" ... "createShareableLink": true
Recommendation

Do not include secrets, regulated data, or private business information unless Manus is approved for that data; consider disabling share links if the API supports it.

What this means

If installed manually from the repository, users must trust the current repository contents rather than a pinned reviewed revision.

Why it was flagged

The documented manual installation pulls code from a GitHub repository without pinning a commit; this is common but makes provenance dependent on the repository at install time.

Skill content
git clone https://github.com/mvanhorn/clawdbot-skill-manus.git ~/.openclaw/skills/manus
Recommendation

Install from the reviewed registry package when possible, or pin and inspect a specific Git commit before manual installation.