Skip to content

Security: dsfsfssdfsd/spark-researcher

Security

SECURITY.md

Security

Spark Researcher runs local project commands and records research/memory artifacts. Treat commands, generated patches, and imported chip logic as privileged local execution.

Launch Boundaries

  • Researcher does not own Telegram ingress.
  • Researcher does not own Spawner mission control.
  • Researcher should be invoked by the operator, Builder, or explicit local CLI flows.
  • Generated code changes should remain reviewable; do not auto-apply unreviewed patches in launch workflows.

Secrets

Never commit:

  • .env, .env.*
  • API keys
  • local project state containing user data
  • generated artifacts that contain private prompts, private code, or credentials

Prefer environment variables or Spark CLI secret storage for provider keys.

Safe Bootstrap

Inspect installer scripts before running them. Do not pipe remote install scripts directly into a shell.

There aren't any published security advisories