Skill flagged — suspicious patterns detected

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

Solvea Chat

v0.5.7

Call Solvea Web App chat API to get AI customer service replies. Use for real customer-service questions. On session boot message ("new session was started")...

0· 278·1 current·1 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Pending
View report →
OpenClawOpenClaw
Suspicious
medium confidence
Purpose & Capability
The skill's name, README and scripts consistently implement a Solvea Web App chat client (chat.py + solvea_client.py + setup.sh). Requiring an API token and agent ID is appropriate for this purpose. However the registry metadata declared 'required env vars: none' while the code enforces SOLVEA_API_KEY and SOLVEA_AGENT_ID and the installer writes a .env—this mismatch is an incoherence.
Instruction Scope
Runtime instructions and templates limit behavior to forwarding user messages to the Solvea API and managing per-user chat IDs in memory/solvea-sessions.json. The installer edits OpenClaw config (openclaw.json) to add bindings/agents which is expected for installing a channel-bound agent. No instructions attempt to read unrelated secrets or exfiltrate data to unexpected endpoints.
Install Mechanism
No remote downloads; the included setup.sh copies the skill into the agent workspace, creates a Python venv and installs two PyPI deps (httpx, python-dotenv). That install mechanism is standard and proportional for a Python CLI skill. setup.sh will write files into the workspace and openclaw.json—normal for a skill installer but requires local config changes.
!
Credentials
The skill legitimately needs SOLVEA_API_KEY (X-Token) and SOLVEA_AGENT_ID; the code will abort if they are missing. The registry metadata incorrectly lists no required env vars or primary credential, which is inconsistent and could mislead users. The installer writes those credentials into a plaintext .env in the skill directory by default—users should be aware of local persistence. Also the default SOLVEA_BASE_URL in code and .env is https://apps.voc.ai while SKILL.md links to https://solvea.cx; this domain mismatch should be verified.
Persistence & Privilege
The skill does not request 'always: true' and follows normal installation behavior. setup.sh will add an agent entry and channel binding to ~/.openclaw/openclaw.json and copy files into the agent workspace, create a .venv, and save .env with credentials. These are expected but are persistent changes to OpenClaw configuration and local disk.
What to consider before installing
This skill's code and installer are coherent with a Solvea chat proxy, but the package metadata incorrectly claims no required env vars—yet the code requires and will persist SOLVEA_API_KEY and SOLVEA_AGENT_ID in a plaintext .env. Before installing: (1) review the setup.sh and scripts (they will modify ~/.openclaw/openclaw.json and copy files into an agent workspace), (2) prefer running setup.sh with --dry-run first, (3) verify the correct Solvea API base URL (apps.voc.ai vs solvea.cx) and that the token you provide is scoped/rotatable, (4) consider creating a dedicated, least-privileged token for this skill and avoid placing other sensitive credentials there, and (5) back up openclaw.json in case you want to revert changes. The primary issue is metadata inconsistency—if you rely on registry declarations, ask the publisher to correct required env var declarations or inspect the code yourself before proceeding.

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

latestvk9711wyvdb9p5ch8zwaw6v2g15832q3p

License

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

Comments