Skill flagged — suspicious patterns detected

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

Brainstorming

v0.1.0

You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requi...

0· 81·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
medium confidence
Purpose & Capability
The skill claims to help with brainstorming and producing design specs and includes a browser-based visual companion and spec-review workflow — the included HTML, helper JS, server, and start/stop scripts are coherent with that purpose. However, the package declares 'instruction-only' and lists no required binaries/env vars while including Node.js server code and shell scripts that clearly require a shell and node to run. That mismatch (not declaring 'node' / shell tooling as required) is a proportion/visibility issue.
Instruction Scope
SKILL.md explicitly instructs the agent to read project files, docs, and recent commits; write design docs into docs/superpowers/specs/ and commit them; run a local visual server and read a .events file for browser interactions; and dispatch a spec-review subagent. Those actions are in-scope for a brainstorming skill, but they grant the skill broad file-system and process control (writing files, running/daemonizing a server, spawning processes). The hard-gate to avoid invoking implementation skills until user approval is clear and appropriate.
!
Install Mechanism
No install spec (no external downloads) is safer, but the shipped files include start-server.sh and server.cjs that require Node.js and a POSIX shell. The skill did not declare required binaries (node, bash, ps/kill/grep/nohup/etc.), which is an incoherence. Starting the server backgrounds processes and writes PID/log files; that behavior is expected for a local companion but has operational risk (orphaned processes, background daemons).
Credentials
The skill requests no credentials or env vars in metadata. The runtime scripts do use optional env hooks (BRAINSTORM_DIR, BRAINSTORM_HOST, BRAINSTORM_URL_HOST, BRAINSTORM_OWNER_PID) but none are sensitive credentials. Be aware the server can bind non-loopback addresses (0.0.0.0) if instructed, which could expose served screens and event files externally — this is not a credential leak from the skill itself but a network exposure risk if misconfigured.
!
Persistence & Privilege
The skill will create session directories (under /tmp or the project .superpowers/ directory), write HTML screens and a .events file, and instruct the agent to write and git-commit spec files to docs/superpowers/specs/. It does not set always:true, but it does expect permission to modify the repository and to spawn/daemonize a local server. That level of file-write and process control is reasonable for this feature but should be consciously allowed by the user.
What to consider before installing
This skill is mostly coherent with a brainstorming + visual companion tool, but before installing: 1) Inspect the included scripts (server.cjs, start-server.sh, stop-server.sh, helper.js) — they will run locally and require Node.js and shell utilities even though the skill metadata doesn't declare those requirements. 2) Expect the skill to create files under /tmp or your project (docs/superpowers/specs/ and .superpowers/brainstorm/), write .events and PID files, and ask the agent to git-commit changes — grant it repo write permissions only if you trust it. 3) When launching the visual companion, ensure it binds only to localhost (avoid 0.0.0.0) unless you intentionally want external access; otherwise you could accidentally expose served mockups or interactive pages. 4) Run the server in an isolated environment (container or throwaway branch) the first time to observe behavior and logs, and confirm your environment has Node.js and the shell utilities the scripts use. 5) If you need stronger assurance, ask the skill author to: declare required binaries (node, bash), remove auto-backgrounding behavior or make it opt-in, and document exactly what files are written and where. If you are uncomfortable with a skill that will modify the repo and spawn a local server, do not install or run it.

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

latestvk97e5fhy6cantq4y5z7j5k47m583dzry

License

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

Comments