Brainstorm

When To Use

Use this skill when the task should run through the native OpenAI Agents Python bundle contract.

OpenAI Agents SDK Compatibility

• Mount this bundle into the Agents SDK Skills sandbox capability.
• Keep scripts/run.sh and scripts/verify.sh deterministic and shell-safe.
• Prefer structured outputs, explicit inputs, and resumable artifacts.

Inputs

• None

Workflow

• discover
• inspect
• plan
• execute
• verify
• summarize
• finalize

Exact Commands

• scripts/run.sh
• scripts/verify.sh

Guardrails

• Use scripts/run.sh and scripts/verify.sh as the declared entrypoints.
• Confine writes to declared output paths.
• Do not finalize before verification passes.
• Keep scripts deterministic, idempotent, and shell-safe.
• Prefer structured outputs that validate against the bundle contract.
• Keep logs trace-friendly with explicit task IDs and outcome messages.
• Preserve compatibility with legacy skill metadata during migration.

Verification

• Run scripts/verify.sh before finalizing any run.

Final Response Checklist

• Verification command completed successfully.
• Outputs are written to declared paths only.
• No secrets were persisted.