No agent in the loop. GitHub issues go in; explicit rules score rejection, recency, engagement, repository activity, and externalizability; the highest-scoring candidates come out.
Two layers: a cheap deterministic radar that continuously searches GitHub for rejected demand, and a smaller curated list of plays worth thinking about harder.
-
The score is intentionally mechanical, not “AI judgment.” It is a radar, not an oracle. Clicking a card goes to the source issue; the suggested product shape is selected by fixed keyword rules.
-__CARDS__
-
+
+
+
__COUNT__raw candidates
+
__CURATED_COUNT__curated plays
+
__REPOS__repos scanned
+
__SEEDS__seed repos
+
__DISCOVERED__discovered repos
+
__UPDATED__last radar · UTC
+
+
+
+
+
curated
+
Promoted opportunities with an actual thesis and smallest next move. Static for now; no scheduled model job burning tokens.
+
+
__CURATED_CARDS__
+
+
+
+
+
raw deterministic radar
+
No model judgment. Seed repos + deterministic GitHub repo discovery → rejected issues → explicit score → ranked candidates.
+
+
Minimum score: __THRESHOLD__. Score = rejection + recency + engagement + externalizability + repository activity/scale/issue surface. Repos are either permanent seeds or discovered from fixed GitHub search queries. Same inputs, same ranking.