Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
IMAP/SMTP Email (Plus)
v1.0.0Read and send email via IMAP/SMTP. Check for new/unread messages, fetch content, search mailboxes, mark as read/unread, and send emails with attachments. Wor...
⭐ 0· 818·7 current·7 all-time
byRoccoon@lroccoon
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The skill's name/description (read/send email via IMAP/SMTP) aligns with the included scripts (imap.js, smtp.js) and README: the code implements the documented functionality. However the package author metadata and registry ownership appear inconsistent (package.json/README claim 'NetEase' but registry owner and _meta.json differ), which suggests repackaging or unclear provenance.
Instruction Scope
Runtime instructions and scripts are scoped to IMAP/SMTP operations: connecting to servers, searching/fetching messages, downloading attachments, and sending email. They read a local .env for credentials and may read/write files for attachments and body files (expected for this functionality). The instructions do not contain explicit exfiltration or unexpected external endpoints beyond standard IMAP/SMTP servers.
Install Mechanism
There is no formal install spec in the registry, but the bundle includes package.json and package-lock.json and expects you to run 'npm install'. That pulls common npm libraries (imap, imap-simple, mailparser, nodemailer, dotenv). This is a moderate-risk pattern—dependencies are normal for an email tool, but the lack of an install spec plus bundled code means the package will write dependencies to disk when you run npm install.
Credentials
The metadata lists no required environment variables or primary credential, yet SKILL.md, setup.sh, and the scripts require many sensitive env vars (IMAP_USER, IMAP_PASS, IMAP_HOST, SMTP_USER, SMTP_PASS, SMTP_HOST, etc.). This mismatch (declaring no env needs while actually requiring account credentials and passwords) is an incoherence and increases risk: the skill will need direct access to email credentials but the registry metadata gives no visibility or prompts about that.
Persistence & Privilege
The skill does not request always:true and does not attempt to modify other skills or global agent configuration. It writes a .env file in the skill folder via setup.sh (local only) and writes attachments to user-specified directories—expected behavior for an email client.
What to consider before installing
This package appears to implement a normal IMAP/SMTP email client, but there are red flags you should consider before installing or entering credentials:
- Credential mismatch: The registry metadata declares no required environment variables, yet the tool requires IMAP_USER/IMAP_PASS and SMTP_USER/SMTP_PASS (sensitive credentials). Treat this as suspicious: verify the author and provenance before supplying credentials.
- Origin/provenance: package.json and README mention 'NetEase' while the registry owner and _meta.json differ—this could be repackaged code. Prefer code from a known/trusted source.
- Install behavior: You must run 'npm install' locally; dependencies are common and expected, but running install will write packages to disk. Inspect the included scripts (imap.js, smtp.js, setup.sh) yourself or in a sandbox before running.
- Minimize exposure: If you decide to test, use an account with no sensitive mail (or an app-specific/test account), and use app-passwords where supported (Gmail app password, authorization codes) rather than your primary account password.
- .env handling: Add the created .env to .gitignore and avoid committing it. Review setup.sh which creates .env to ensure it writes only intended values.
- If you cannot verify the publisher, do not provide real credentials. Consider running the tools in a restricted environment or sandbox and review network connections/logs while testing.
If you want, I can summarize exactly where the code reads/writes credentials and which files to inspect before running, or help craft safer test steps (sandboxed run, throwaway account, or manual credential parsing).Like a lobster shell, security has layers — review code before you run it.
latestvk974nv3bga08bzvv3bvsrff2a981p4wd
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
