Skill Federation.
How it works Benchmark Privacy Security View on GitHub →

Privacy Policy

Last updated: 2026

1. What Crosses the Boundary

Skill Federation is private by design. When your agent needs a skill, only an abstract wish is sent to the search service:

  • A one-line description of the capability needed
  • A few vocabulary-varied paraphrases of that description
  • 1–5 keywords

That is the complete set of data used to search the catalog.

2. What Never Crosses

The following are never transmitted to or stored by the Service:

  • Your source code, code snippets, or file contents of any kind
  • Your plan, brief, prompts, or reasoning traces
  • Your agent's outputs
  • File paths, directory structures, or repository names
  • Project names, product names, or internal tool names

Your work never leaves your machine. Retrieved skills are installed and run locally.

3. How Search Works

The abstract wish is embedded and compared against a vetted catalog of publicly sourced skills. The Service returns the best matches — each with its license class, provenance, source, and any security flags — for you to review and approve.

Catalog skills are ingested from public repositories, copied, deduped, and scanned before promotion. This ingestion pipeline operates on public data, not on your private inputs.

4. Data Storage

Service infrastructure is hosted in AWS US regions with encryption at rest and TLS in transit. We store only the minimum operational data needed to run the search service and, for team customers, workspace membership and access controls.

5. Third Parties

Catalog vetting relies on independent security scanners, including Cisco AI Defense and NVIDIA SkillSpector, which operate on catalog candidates (public skills), not on your private inputs. Embedding and matching may use model providers under endpoints that do not train on submitted data.

We do not sell your data. We do not use third-party advertising or cross-site tracking.

6. Your Rights

Because the finder runs locally and free use requires no account, there is little personal data to manage. For team and enterprise accounts you may delete your account and associated credentials, and export any submission history retained for your workspace.

7. Cookies & Tracking

The public marketing site uses no third-party analytics, advertising trackers, pixel tags, or fingerprinting. Authenticated dashboards use an essential session cookie only.

8. Changes to This Policy

We will provide reasonable notice before making material changes to this policy. Continued use of the Service after the notice period constitutes acceptance of the updated policy.

Questions? Contact us at privacy@qstarlabs.io.

Skill Federation github.com/skill-federation · Privacy · Terms © 2026 Qstar Labs