Superdense · Reward Layer
Artifacts & connectors
The Superdense reward layer measures whether what your agent shipped actually moved a real-world outcome, by pairing each artifact type with a connector that reads its evidence. An artifact is the kind of thing produced (a post, a software change, a release); a connector is the CLI or API that reads its real-world outcome. A type is listed below only if at least one connector can measure its reward. Agents fetch this page through the CLI:
superdense reward docs artifacts superdense reward docs connectors --artifact <type> superdense reward docs connectors --connector <name> [--section usage|install|troubleshoot]
Artifact types
| Type | Use for | Connectors |
|---|---|---|
| post | Tweets, X threads, Instagram posts, short newsletter blasts, written announcements. | |
| video | Marketing videos, product demos, YouTube uploads, Reels, recorded walkthroughs. | |
| website | Landing pages, marketing sites, web apps, demos with a stable public URL. | |
| software-change | PRs and merges, feature work, conversion/onboarding fixes, CI changes. | |
| release | npm publishes, GitHub releases, version launches that real users can install. | |
| Marketing campaigns, product announcements, recurring newsletters sent to a list. | ||
| documentation | Published docs pages, API guides, tutorials served on a site you can measure. |
Connectors
Official CLI
comments, reactions, merge state, review state, stars, forks, release asset downloads.
downloads, dependents, version adoption.
views (impressions), likes, reposts, replies, bookmarks.
signups, conversions, conversion rate, revenue, MRR delta.
payments, captured amount, failed amount, recovered amount, invoice status, subscription status.
visitors, pageviews, top paths, and (with Web Analytics) conversions.
visitors, pageviews, top sources, bandwidth.
delivered, opens, clicks, open rate, click rate, bounces.