I write about papers and books I'm reading, and things I'm building, on my blog here.
Reach me on X
- REA: Reverse Engineer Anything
- One CLI and MCP server that lets coding agents inspect any app, understand how it works down to the binary level, and recreate features for your own product.
- Preference - Research-driven data infrastructure for trading agents.
- gstack
- #1 outside contributor; project-local /browse install hardening and setup fixes.
- Model Context Protocol
- TypeScript SDK packaging fix for CommonJS validator exports and Ajv.
- DeepSeek Agents
- DeepSeek integration guides.
- Qwen Code
- Config validation for fractional session and tool-call limits, max-token continuation rollback, MCP read-only auto-approval trust enforcement, and malformed tool result display resilience.
- Hermes
- Cron scheduler double-execution prevention and credential-pool failover fix for provider auto-detection.
- x-JEPA
- Correctness fixes across the JEPA training and planning stack: Wasserstein midpoint targets, unimodal Beta policies, goal/state token alignment, SIGReg dtype handling, and discrete planning action indices.
- OpenClaw
- Shared phone identity canonicalization, Signal/iMessage routing fixes, WhatsApp normalization coverage, and stale-chunk recovery unblock when the document probe hangs.
- RepoPrompt
- Context Builder model selection, agent-mode and tool-card reliability, workspace model drift cleanup, test harness hardening against hangs and waiter leaks, and hosted-app/XCTest sharding.
- Mooncake
- RDMA completion resource validation, cross-thread SHM relocation mapping lifecycle, and NVMe-oF task batch rejection with cuFile completion correlation.
- Vibe-Trading
- Codex OAuth init/default-model fixes and CLI resume prompt input preservation.
- Magpie
- Fallible JSON parsing across FFI, runtime, codegen, semantic analysis, and web lowering.
- HarnessGym
- Benchmark correctness fixes for result parsing, activation reporting, and artifact quarantine.
