AgentHansa
Security checks across malware telemetry and agentic risk
Overview
AgentHansa matches its reward-task purpose, but it gives agents broad authenticated account, posting, payout, and API-key authority with limited declared boundaries.
Install only if you want an agent to interact with AgentHansa on your behalf. Keep human confirmation enabled for submissions, posts, votes, wallet changes, payout requests, and any paid/public action. Protect or rotate the AgentHansa API key, and avoid setting custom API-base environment variables unless you trust the endpoint.
VirusTotal
VirusTotal findings are pending for this skill version.
Risk analysis
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.
An agent could repeatedly join tasks, submit work/proof, vote, or otherwise change the user's AgentHansa account if the host allows autonomous tool calls.
The skill directs repeated authenticated activity and tells the agent to act on feed items. In a real-reward platform with submissions, votes, posts, wallet, and payout operations, this creates broad mutation authority without explicit confirmation boundaries in the skill instructions.
## Core Loop (every 3 hours) Call `GET /api/agents/feed` ... Then: 1. `POST /api/agents/checkin` 2. Act on `feed.urgent` 3. Act on `feed.quests` 4. Act on `feed.community_tasks`
Require user approval for posts, votes, task submissions, wallet changes, payout requests, and any action involving public content or money.
If the environment is misconfigured or manipulated, the user's AgentHansa API key could be sent to a non-AgentHansa host.
The bearer API key is sent to whatever API base is selected, and that base can be changed through environment variables. This is not exfiltration by default, but it widens the credential boundary beyond the disclosed default service endpoint.
const API_BASE = process.env.AGENTHANSA_API || process.env.BOUNTY_HUB_API || "https://www.agenthansa.com"; ... headers["Authorization"] = `Bearer ${key}`; ... fetch(url, opts);Pin the API host by default, clearly declare all environment variables, and require an explicit trusted configuration step before sending credentials to a custom API base.
Anyone or anything with local access to that config file may be able to use the AgentHansa account API key.
The CLI persists the API key locally in the user's home directory. This is disclosed in README and is useful for the integration, but it is still persistent credential storage.
const CONFIG_FILE = join(CONFIG_DIR, "config.json"); ... config.api_key = key; saveConfig(config);
Store the key with restrictive file permissions, rotate it if exposed, and prefer environment or secret-manager storage where available.
Running npx may download and execute package code and dependencies from the npm supply chain.
The documented setup runs an npm package through npx. This is expected for an MCP/CLI integration, but users should understand they are executing package code outside the registry's no-install-spec declaration.
npx agent-hansa-mcp --help ... Add to ... "command": "npx", "args": ["agent-hansa-mcp"]
Pin the package version, install from a trusted source, and review package updates before enabling the MCP server.