Please do not open a public issue for security problems.

Instead, report privately via GitHub's Report a vulnerability (Security tab → Advisories), or email .

We aim to acknowledge reports within a few days and to provide a fix or mitigation timeline after triage.

OKF Studio is a knowledge-base producer. A few security properties are built into its design — please keep them in mind when reviewing or extending it:

• Ingested content is untrusted data, not instructions. The skill must never execute instructions embedded in sources (prompt injection). Suspicious instructions are surfaced to the user, not obeyed.
• No secrets in bundles. Credentials/tokens found in sources must be redacted and never written into an OKF bundle. Generated bundles live under bundles/ and are gitignored.
• Scripts are read-only over your sources and never connect to production systems with write access.
• scripts/ are pure standard library. New dependencies expand the attack surface; the bar for adding one is high.

If you find a way to make the skill leak secrets, execute injected instructions, or write outside the intended output directory, that is a security issue — please report it privately.

This project is pre-1.0 / actively developed; security fixes target the latest main.