Brainstorm

Generate ideas fast. Adapt depth and structure to what the user actually needs.

Security Scan

This is an instruction-only brainstorming skill whose files, runtime instructions, and requirements are consistent with its stated purpose and request no extra privileges or credentials.

Purpose & Capability

The name/description (generate ideas, adapt depth/structure) match the SKILL.md, modes.md, and prompts.md. The skill requests no binaries, env vars, or config paths, which is appropriate for an idea-generation helper. Minor metadata inconsistency: _meta.json.ownerId is different from the registry metadata ownerId listed in the package header — likely a benign metadata mismatch but worth noting.

Instruction Scope

SKILL.md instructs only on how to generate outputs, when to ask a single clarifying question, and how to structure results. It does not instruct the agent to read files, access environment variables, contact external endpoints, or collect unrelated data. It references related modes/skills ('diverge', 'loop') that are not included; this is a usability/clarity issue, not a security risk.

Install Mechanism

No install spec and no code files — instruction-only. Nothing is downloaded or written to disk by the skill itself.

Credentials

The skill declares no required environment variables, credentials, or config paths, and the instructions do not reference any secrets. The access requested is minimal and proportional to the purpose.

Persistence & Privilege

The skill does not request always:true and uses default invocation settings. It does not modify other skills or system settings. Autonomy settings are the platform defaults and are not a concern here.

Assessment

This skill appears safe and coherent: it only contains instructions for producing brainstorming outputs and asks for no credentials or installs. Two small things to check before installing: (1) the _meta.json ownerId differs from the registry ownerId — confirm the package provenance if you care about strict ownership tracking; (2) the SKILL.md refers to related modes/skills ('diverge', 'loop') that aren't bundled here — expect broken cross-references or missing behavior if you rely on those. Otherwise you can use this skill without granting extra privileges or secrets.

Default: Ideas First

When user asks for ideas, generate immediately. Don't interrogate first.

If context is missing, the ideas will show it — user will clarify naturally.

Adapt to the Request

Signal	Response
"Give me 5 ideas"	5 ideas, simple list
"I'm stuck on X"	3-5 angles to unstick
"Brainstorm names"	10-20 options, varied styles
"Help me think through X"	Explore, ask clarifying Qs
"What are my options?"	Structured comparison

Match the energy. Quick request → quick response. Deep request → go deep.

Only Ask If

• Request is genuinely ambiguous (not just incomplete)
• You're about to invest significant effort
• Constraints would completely change the output

One question max. Not five.

Structure When It Helps

Don't force categories. Use structure only when it clarifies:

• Comparisons: Table with trade-offs
• Many ideas: Group by theme
• Recommendations: Highlight top pick with reasoning

Flat list is fine for most requests.

After First Round

If user wants more: go deeper, wider, or different direction. If user is silent: they got what they needed. Don't push.

Check modes.md for different brainstorm approaches. Check prompts.md for creative expansion techniques.

---

Related: For multi-perspective exploration, see diverge. For iteration, see loop.