dsh-csp-runtime
Run the following command in DeepSeek Harness:
dsh plugin install helibeiqi/dsh-csp-runtime
Paste the following prompt into your AI chat to install this plugin:
Run `dsh plugin install helibeiqi/dsh-csp-runtime` in DeepSeek Harness to install this plugin; source code is available at https://github.com/helibeiqi/dsh-csp-runtime .
About this plugin
In multi-agent pipelines, an agent's reasoning is often ephemeral: where the goal sits, what beliefs drove a decision, what remains uncertain, and how the next step should be handed off—all of this vanishes the moment the process exits, forcing the receiving agent to re-derive everything from scratch. dsh-csp-runtime introduces the Cognitive State Protocol (CSP), packaging goal, belief, plan, decision, uncertainty, and handoff into a single pure-JSON snapshot so that reasoning becomes a first-class, serializable, persistable, and migratable artifact.
Its positioning is an interoperability layer, not a replacement for event sourcing. DSH continues to record sessions in an event-sourced fashion internally; this plugin simply exports that stream into a generic CSP format so that CrewAI, LangGraph, or any JSON consumer can import and continue the work without understanding the DSH kernel. The most original design is the handoff token: three-state semantics (transferable, final, blocked) explicitly declare whether a state can be picked up, and in what manner, drawing a clear line between a cognitive snapshot and a plain session log.
It suits teams building multi-agent pipelines on DSH who need to export intermediate reasoning to heterogeneous frameworks; orchestrators who require explicit task handoff and blocking markers between agents; and engineers who want an auditable, migratable protocol snapshot for agent cognition without altering the underlying event-sourcing architecture. Note that this is an M1 Developer Preview; lock your version before production use.
Use Cases
- Export mid-pipeline reasoning to heterogeneous frameworks like CrewAI or LangGraph
- Declare handoff mode (transferable, final, blocked) so the next agent knows exactly how to proceed
- Export DSH event streams to a generic CSP JSON for non-DSH consumers to import without touching the DSH kernel
Best For
- Teams building multi-agent pipelines on DSH who need to share cognitive state across frameworks
- Orchestrators who require explicit task handoff and blocking markers between agents
- Engineers who want an auditable, migratable protocol snapshot without modifying the underlying event-sourcing architecture
Related Plugins
Traceable, searchable cross-session memory for AI agents that turns conversation knowledge into a typed knowledge graph and recalls relevant subgraphs instead of replaying full history, natively integrated with DeepSeek Harness.
Gives DSH AI cross-session long-term memory, to-do and skill management, plus multi-session orchestration, external AI delegation, and an infinite canvas that grows with you.
dsh-mnemon is a three-tier, pluggable, Agent-driven memory system for DeepSeek Harness, combining Runtime memory, Project Documents, and replaceable Memory Spaces with nine long-term providers.
An opinionated, zero-infrastructure file-based memory protocol: human-readable markdown files plus a bounded always-loaded index, curation discipline, and no database or embeddings.