Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Brainstorming
v0.1.0You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requi...
⭐ 0· 147·0 current·0 all-time
by@liuhe12
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 (brainstorming + visual companion) match the included files: a small local Node HTTP/WebSocket server, helper client script, and UI template. However the runtime expects Node and performs git commits and file writes to the project (docs/superpowers/specs/...), yet the skill declares no required binaries or environment — an omission that should be clarified.
Instruction Scope
SKILL.md instructs the agent to 'check files, docs, recent commits', write spec files into docs/superpowers/specs/YYYY-... and commit them, start a local server, watch a session directory, and read/write a .events file. Those actions involve scanning repository history, reading project files, writing files, and performing commits — all beyond a purely conversational helper and require explicit user consent and appropriate environment safeguards.
Install Mechanism
There is no install spec (instruction-only), which limits automatic installation risk. But the bundle includes executable scripts that run node server.cjs and shell scripts that background processes; Node.js and a shell are implicit runtime requirements that are not declared. That mismatch is worth noting.
Credentials
The skill requests no env vars, yet the scripts read/process env vars (BRAINSTORM_DIR, BRAINSTORM_HOST, BRAINSTORM_URL_HOST, BRAINSTORM_OWNER_PID) and will create files under /tmp or under the project (.superpowers/brainstorm/). It also expects to be able to commit to the repo (git) and to access recent commits. File-system write/read and repo commit privileges are significant and are not called out in the declared requirements.
Persistence & Privilege
The skill launches a background HTTP/WebSocket server that serves files from a session directory and can be configured to bind non-loopback interfaces (0.0.0.0). That introduces a network-exposure risk (serving project files or mockups to the network). 'always' is false, but since the agent can invoke the skill autonomously, it could start/stop the server during a session — ensure this is acceptable and that the server is run in a safe environment.
What to consider before installing
This skill is functionally coherent with a brainstorming + visual companion workflow, but it runs a local Node-based server, writes files into your project, and expects to create/commit spec files. Before installing or running: 1) Confirm you trust the skill owner and review the included scripts (server.cjs, start/stop scripts, helper.js, frame-template). 2) Ensure Node.js and a shell are available and that you want an AI to write files and run commits in your repo (it will write to docs/superpowers/specs/ and may run git). 3) Prefer using --project-dir so files are placed under a controlled path, add .superpowers/ to .gitignore, and run the server in an isolated environment (container or VM) if you are worried about network exposure. 4) If you must run it on a remote/container host, avoid binding to 0.0.0.0 unless you intend to expose the UI; prefer localhost. 5) If you need a more conservative install, ask the skill author to declare required binaries (node, git), document exactly what is committed, and provide an opt-in confirmation step before any commits or server starts.Like a lobster shell, security has layers — review code before you run it.
latestvk9771w95ram3259xh0fpzkhwjh8374hb
License
MIT-0
Free to use, modify, and redistribute. No attribution required.