Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Heleni WhatsApp
v1.1.0Complete WhatsApp management for OpenClaw agents: per-conversation memory (groups + DMs), unanswered message tracking, loop prevention, and multi-PA coordina...
⭐ 0· 56·1 current·1 all-time
byNetanel Abergel@netanel-abergel
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The name/description promise 'complete WhatsApp management' (unanswered tracking, loop prevention, multi-PA coordination) but the skill only provides local filesystem conventions and shell helper functions for reading/writing context and inbox JSON. There is no code or instructions to connect to WhatsApp (API, webhook, or gateway). That makes the claimed capability larger than what is provided — plausible as a companion/helper for an existing integration, but the README does not say that explicitly.
Instruction Scope
The SKILL.md tells the agent to source a local CONTEXT_FILE (/opt/.../.context) which can set arbitrary environment variables or execute shell code — sourcing is effectively code execution. It also references runtime variables like OWNER_PHONE, JID_CORE_TEAM, INBOX_FILE that are not declared or validated. The instructions mandate writing a context file for every DM (no exceptions), which means persistent logging of conversation contents. Overall the instructions reach outside a narrow helper role (they execute/shell-source state and create many files).
Install Mechanism
Instruction-only skill with no install steps, no third-party downloads, and no code files. That minimizes supply-chain risk; nothing is written or executed at install time by the registry itself.
Credentials
The skill declares no required env vars but the instructions expect variables like OWNER_PHONE, JID_CORE_TEAM, and INBOX_FILE when sourcing .context. Sourcing an externally provided context file can expose secrets or execute arbitrary commands. The skill asks the agent to create/read files under $HOME/.openclaw (or /opt in one snippet) but gives no guidance on what sensitive values might be stored there. Requesting/using undeclared env/state is disproportionate to a simple memory helper.
Persistence & Privilege
The skill does not set always:true and doesn't modify other skills. However, it mandates persistent storage of all DM context files and an inbox/pending.json. That persistence is expected for a memory-tracking skill, but because it is unconditional ('No exceptions. Every DM = a context file'), it can accumulate potentially sensitive data on disk; review where those files are stored and who can read them.
What to consider before installing
Before installing, consider these points:
- This skill is instruction-only: it will not itself connect to WhatsApp. It's a local file-based memory and inbox convention that an agent/integration would need to implement separately. If you expected automated WhatsApp integration, this skill alone does not provide it.
- The SKILL.md directs the agent to 'source' a .context file (/opt/.../.context). Sourcing runs shell code and can execute arbitrary commands or inject environment variables. Only allow this if you control and have audited the .context file contents.
- The instructions reference undeclared environment variables (OWNER_PHONE, JID_CORE_TEAM, INBOX_FILE) and will create persistent files under $HOME/.openclaw (and earlier references to /opt). Decide where those files should live and who can read them; they may contain sensitive conversation content.
- The skill mandates writing a context file after every DM, which can lead to broad local storage of message contents — consider data retention, encryption, and access control.
- If you want to proceed: (1) inspect the .context file the agent would source and never allow untrusted .context files; (2) confirm or change the storage path to a safe, access-controlled directory; (3) require explicit integration code/credentials for any WhatsApp connectivity so the system behavior is clear.
Additional information that would raise confidence: an explicit statement that this is only a memory helper (not a connector), the canonical location and format for .context with an example, and evidence of an official author/homepage or code that performs WhatsApp API integration. If those are provided, reassess from 'suspicious' toward 'benign.'Like a lobster shell, security has layers — review code before you run it.
latestvk975vtq2kbkk80dtyp0w7khjwh84acvt
License
MIT-0
Free to use, modify, and redistribute. No attribution required.