Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

iClawd Email

v1.0.0

Get a real email address for your AI agent. Create an inbox, send and receive email, with optional PGP encryption and DID verified identity.

1· 41·0 current·0 all-time
byBadr@karimibadr
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
medium confidence
Purpose & Capability
The name and description match the instructions: the skill talks to https://iclawd.email/mcp to create/send/receive mail. That functionality reasonably explains HTTP API usage and webhooks. However the skill's runtime docs clearly require an API key and local config storage even though the registry metadata declares no required credentials or config paths — this is an incoherence.
!
Instruction Scope
SKILL.md instructs agents to store and reuse an API key, set webhooks to arbitrary URLs, auto-create addresses, and send external mail (including signing up for services). It also tells agents to save the API key to ~/.iclawd/config.md and not to include that key in outgoing mail. Those file-path and secret-handling instructions are outside what the registry declared and grant the skill the ability to read/write a local config and to forward inbound messages to external endpoints (webhooks), which are higher-risk operations.
Install Mechanism
This is an instruction-only skill with no install spec, no downloaded artifacts, and no code files to execute. That minimizes direct install-time risk; the runtime risk is from the external API and the actions the agent is instructed to take.
!
Credentials
Although the registry lists no required environment variables or primary credential, the SKILL.md repeatedly uses an API key (Authorization: Bearer $API_KEY) and instructs saving it to ~/.iclawd/config.md. Requesting a persistent API key and local storage without declaring it in metadata is an inconsistency and increases the chance of accidental secret exposure. Webhook configuration also allows exfiltration of incoming mail to arbitrary endpoints.
Persistence & Privilege
The skill does not request always:true and does not appear to modify other skills. It instructs the agent to create and store a per-agent API key in a local config file and to register webhooks. Persisting a config file and webhook registrations is expected for a mailbox service, but combined with the undeclared API key and webhook flexibility it raises a moderate privilege/risk surface.
What to consider before installing
This skill appears to do what it says (give an agent an @iclawd.email inbox) but there are important mismatches and risks to consider before installing: - The SKILL.md expects an API key and directs saving it to ~/.iclawd/config.md, yet the registry metadata declares no required credentials or config paths. Ask the publisher to clarify how credentials are issued, stored, rotated, and what exact permissions the API key has. Do not assume the key is low-privilege. - Webhooks can send inbound emails to any URL you configure. Only allow webhooks to endpoints you control and trust; otherwise incoming mail (possibly containing sensitive content) could be forwarded externally. - The skill permits automated sending (including signups and replies). Decide whether the agent should be allowed to send external emails autonomously or require owner approval for any external recipient or attachments. - Verify the domain and service legitimacy (https://iclawd.email). Because the skill's source/homepage are 'unknown' or not verified in the registry metadata, confirm the provider identity and review privacy/retention policies before handing over real messages. - Ask the author to update the registry metadata to list the API key as a required credential and to declare the config path(s). Prefer skills that declare required env vars/config paths in metadata so the platform can surface permission prompts to users. If you proceed, restrict the agent: require explicit owner confirmation for webhook creation, external sends beyond occasional signups, and any operation that would store or forward owner-sensitive content.

Like a lobster shell, security has layers — review code before you run it.

latestvk97drh3vkdbbqxahga9mnv4rhn83pw12

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Comments