10 projects
swarph-cli
The `swarph` binary — multi-LLM CLI + mesh-gateway integration: multi-provider `swarph spawn` (claude/codex/antigravity per cell.provider via a ProviderMembrane + subprocess billing-scrub), interactive `swarph init`, `swarph mesh` (send/inbox/register) + inbox sidecar, `assisted_memory` (git-backed durable memory), session import, watchdog. v0.9.5: foolproof Windows auto-relaunch (genuine-WT detection via process ancestry, not WT_SESSION).
swarph-mesh
Model-agnostic Python substrate for the swarph-mesh ecosystem — the graph protocol third-party CLIs plug into. v0.7.9 adds the GrokCLIAdapter (provider 'grok-cli'): the $0 SUBSCRIPTION Grok lane via firejail-sandboxed `grok --prompt-file` on a SuperGrok-Heavy OIDC session token (sibling of the metered 'grok' adapter), with prompt-file 0600 + memory-store blacklist + endpoint-redirect scrub hardening.
dispatch-relay
Provider-agnostic LLM dispatch layer: 3 injected seams (config / usage / dispatch) + a pure cost model. Relays usage to a sink rather than tracking it.
swarph-shared
Shared substrate primitives for swarph-mesh ecosystem (caller convention, subprocess env scrubbing, JSON-mode harness, peer-name registry, cell.yaml schema). v0.3.2 bundles the adversarial-sweep hardening: subprocess_env billing-redirect scrub, string-aware JSON brace extraction, and peer_registry empty-token/malformed-shape/regex-unification fixes.
fisherrao
Fisher-Rao geometry on multiple statistical manifolds — multivariate Gaussian (Calvo-Oller) + Beta family (KL, geodesic, curvature, MoM fit). Pure numpy+scipy.
langgraph-genai-bridge
Native Google GenAI SDK integration for LangGraph with context caching, tool auto-conversion, and bidirectional message translation.
hodgex
Hodge Laplacian + simplicial complex + 3-way Hodge decomposition on graphs. Pure scipy.sparse, 2-simplex (triangle) scope.
diebold-yilmaz
Diebold-Yilmaz (2012) connectedness — generalized FEVD + total/directional/net spillover indices. Pure numpy+scipy.
tailcor
TailCoR — non-asymptotic, distribution-free decomposition of linear correlation vs non-linear tail contagion (Ricci, Tristani, Vergote 2023). Pure numpy+scipy.
phawkes
Pure-Python multivariate Hawkes process simulation and MLE (exponential kernels), no C++ deps
