dsh-session-handoff
Run the following command in DeepSeek Harness:
dsh plugin install snow-The/dsh-session-handoff
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install snow-The/dsh-session-handoff in DeepSeek Harness to install; full source at https://github.com/snow-The/dsh-session-handoff
About this plugin
Long sessions hitting context limits, repeated auto-compaction, and lost progress on every new session — these are the daily headaches of heavy DeepSeek Harness users. dsh-session-handoff was built to solve exactly that: a zero-dependency core with no build step, packaging the full workflow from safe handoff to proactive compaction into independently callable tools.
Five modules cover the ground. Handoff parses the current session into a structured Markdown document that a fresh session can load and continue from seamlessly; ACP compresses consumed history before the window fills, keeping the original text as a searchable hidden layer; Session management offers trash-style archive, restore, and permanent delete with recoverable mistakes; Model Routes let a single model id switch freely across official API, Volcano Ark, custom endpoints, and vision-toolkit wrappers, with automatic failover on quota or timeout errors; the Web GUI maps all of this into a drag-to-reorder panel, slider controls, and one-click export buttons — no command memorization needed.
It is best suited for users running long-horizon agent workflows, juggling parallel sessions, or sharing one model across multiple vendors on Harness, and for developers who want to fine-tune compaction thresholds without editing YAML. If you only have short conversations, the payoff is limited; but once a session runs into tens of thousands of lines, it quickly becomes essential.
Use Cases
- Safely hand off a 170k-line session to a fresh one and continue seamlessly
- Auto-failover between official API and Volcano Ark routes on quota or timeout
- Tune compaction thresholds and route priority in real time without editing YAML
- Resume a previous session in a new context by loading the handoff document
Best For
- Heavy DeepSeek Harness users running long-horizon agent workflows
- Dev teams sharing a single model id across multiple vendors
- Developers who need fine-grained control over context window behavior and compaction strategy
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.