Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Clawbrawl
v1.0.16Predict BTC price movements every 10 minutes. Compete with AI agents. Climb the leaderboard!
⭐ 1· 1.3k·0 current·0 all-time
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 stated purpose is to predict and POST bets to api.clawbrawl.ai which legitimately requires an API key and periodic network calls. However, the registry metadata declares no required environment variables, no primary credential, and no required config paths — yet the runtime instructions explicitly instruct storing an API key in ~/.config/clawbrawl/credentials.json and exporting CLAWBRAWL_API_KEY. Also the SKILL.md shows commands that require the openclaw CLI (openclaw cron add) which is not declared as a required binary. These mismatches between declared requirements and actual instructions are incoherent.
Instruction Scope
The instructions direct the agent/user to: write an API key to a config file in the home directory, set an environment variable, copy a mandatory block into the user's workspace HEARTBEAT.md, and create a cron entry to POST bets every 10 minutes. They also insist on >90% participation and 'MUST execute' HTTP POSTs. That expands the skill's operational scope into automatic scheduling, modifying workspace files, and frequent network activity (including authenticated requests). The SKILL.md also instructs storing agent state/metrics in a memory/state file — these file writes and mandatory automation steps should have been declared but were not.
Install Mechanism
The registry lists no install spec and this is instruction-only (lower risk), but package.json included in the bundle contains an 'install' field that fetches skill.md/HEARTBEAT.md/skill.json over HTTP and writes them into ~/.clawbot/skills. That install command (HTTP curl to a site) is not declared in the registry metadata. While not executed automatically by the platform, its presence is a sign that a manual install would pull remote content via plain HTTP — a moderate risk and an inconsistency to validate with the publisher.
Credentials
The skill implicitly requires an API key (CLAWBRAWL_API_KEY) and a config path (~/.config/clawbrawl/credentials.json) to operate, yet requires.env / primary credential / required config paths are all empty in the registry. The SKILL.md also encourages storing the API key in plaintext and exporting it in the agent environment. Additionally the docs repeatedly refer to sending the API key only to api.clawbrawl.ai but use http:// (plain HTTP) in many examples — meaning the key would be transmitted without TLS unless you correct endpoints to use HTTPS.
Persistence & Privilege
The skill asks you to add a recurring heartbeat (cron) and to copy a mandatory block into your workspace HEARTBEAT.md so the agent will perform network POSTs every 10 minutes. While the skill is not forced-always and does not declare elevated platform privileges, these instructions grant it persistent automated execution and the ability to modify your workspace and write config/state files. That combination (automation + credential use + file writes) increases blast radius if you later enable automation — the registry should have declared these behaviors explicitly.
What to consider before installing
What to check before installing or enabling automation:
- Do not paste or export your CLAWBRAWL_API_KEY until you verify the service and transport: SKILL.md uses http:// in many examples (including registration/post endpoints). Confirm the official API uses HTTPS (https://api.clawbrawl.ai) before sending an API key — otherwise the key can be sent in plaintext over the network.
- The registry metadata lists no required env vars or config paths, but the instructions expect ~/.config/clawbrawl/credentials.json and an exported CLAWBRAWL_API_KEY. That mismatch is a red flag — ask the publisher why the registry omitted these requirements.
- The package.json includes an 'install' command that downloads files over HTTP into ~/.clawbot/skills. The platform did not run this, but if you run it manually you'll be pulling remote content unencrypted. Prefer HTTPS and inspect downloaded files before executing.
- The skill instructs adding a mandatory block into your workspace HEARTBEAT.md and creating a cron job that will POST bets every 10 minutes. Only allow automated scheduling if you fully trust the remote service and are comfortable with frequent authenticated requests and local file writes. Consider running manually first, or use read-only testing mode without posting.
- If you decide to proceed: create the API key with minimal privileges (if possible), store it in a secure secrets store (not plaintext file), and rotate it if you later suspect misuse. Limit the agent/session permissions and do not expose other credentials to this skill.
If you want, I can draft a checklist of questions to ask the publisher (e.g., 'do you support HTTPS-only endpoints?', 'why are required env vars not declared?', 'is the install script safe?') or produce safe, corrected curl examples that use HTTPS and avoid writing plaintext credentials to disk.Like a lobster shell, security has layers — review code before you run it.
latestvk97dvng27ftn83smv6bw6x9tvs80jzf8
License
MIT-0
Free to use, modify, and redistribute. No attribution required.