Back to search
Top Contributors
Cognithor - Agent OS: Local-first autonomous agent operating system. 16 LLM providers, 17 channels, 112+ MCP tools, 5-tier memory, A2A protocol, knowledge vault, voice, browser automation, Computer-use, self-healing, self-improving. Python 3.12+, Apache 2.0.
Stars
70
Forks
10
Watchers
70
Open Issues
0
Overall repository health assessment
No package.json found
This might not be a Node.js project
feat(arc): path shortening for click solutions + incremental DFS integration
91afbadView on GitHubfeat(arc): incremental Click-DFS for deep click puzzles (LP85 L2)
dfde1e0View on GitHubfeat(arc): LP85 L1 solved! Tuned scanner: diff>10 threshold, 15% tolerance, max 4 groups
7c402b1View on GitHubrelease: v0.73.0 — ARC-AGI-3: 7/25 Games Solved, 18 Levels
3960093View on GitHubfeat(arc): try click/interact at each DFS position for mixed games
bba0e79View on GitHubfeat(arc): smart action ordering in DFS — avoid immediate reversal
1d2ba4cView on GitHubfeat(arc): DFS path shortening — remove redundant steps after solving
aecb2b8View on GitHubfix(arc): double-step for delayed-render games, _replay_to helper
570b334View on GitHubfeat(arc): KeyboardSolver — incremental DFS for keyboard games
12434daView on GitHubfix(arc): always try BFS before sim-A* (fixes LP85 Level 0)
37e5906View on GitHubfix(arc): detect false positives — verify levels_completed before counting solved
bbb0f95View on GitHubfix(arc): debug L3 state range, BFS fallback heuristic, dim range logging
405c45eView on GitHubfix(arc): multi-color markers, BFS fallback when no markers, deeper search limits
e679d3cView on GitHubfeat(arc): simulation A* — real env clicks with height-based dedup
10afba5View on GitHubfeat(arc): full height-space A* with 5D search, pump-then-trigger architecture
f5bf4c4View on GitHub