A 267-skill research graph in pure markdown — 51 research operations built from 216 single-purpose steps, composed in any order with explicit backtracking. One npx install, no runtime, no MCP bindings. The AI is the researcher; you set the direction.
A DeepSeek Harness (dsh) plugin — turns dsh web into a SillyTavern-style roleplay frontend: character cards, prompt presets, lorebooks, personas, BM25 long-term memory, world-state deltas, and rollback-able floor operations, all on the dsh agent runtime.
Make destructive AI-agent actions reversible by default — quarantine + audit + human escalation for rm/git destructive operations. Reliability infrastructure, not a sandbox.
A remote workspace plugin for DeepSeek Harness (DSH). It manages SSH hosts inside DSH, opens remote terminals, lets you browse and edit remote files and inspect remote Git, and adds remote directories as workspaces, so the Agent's file operations and commands run on the remote machine.
Behavior-drift reports for DeepSeek Harness agents: compare two session logs and see exactly how behavior changed — tool sequence, file operations, reasoning volume, timing, retries — zero dependencies, CI-ready.
DeepSeek Harness watchdog plugin: detects truly stalled agent turns (never killing in-progress tasks — in-flight operations are exempt), nudges/terminates only on real silence, records every event to JSONL with a loopback status route
Model Capability Facade for DeepSeek Harness: declare one semantic capability, get typed model-facing operations that run a deterministic pipeline of existing tools.
