dsh-premise-guard-cn
Run the following command in DeepSeek Harness:
dsh plugin install 111111AAA111111/dsh-premise-guard-cn
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install 111111AAA111111/dsh-premise-guard-cn in your terminal to install this plugin; the source is available at https://github.com/111111AAA111111/dsh-premise-guard-cn
About this plugin
A long DSH session for Chinese novel or long-form writing easily spans dozens of turns. Once the context is compacted, the model has no clue what it just dropped—key premise anchors like the sect wiped out on a single night or the ninth heaven no one has ever truly reached can vanish from the summary, and every chapter after quietly drifts off course with no one the wiser.
dsh-premise-guard-cn exists to catch that drift. After every compaction/summary event it diffs the compressed span against the summary text, auto-extracting six categories of Chinese anchors (quoted phrases, dotted term chains, IDs, relative paths, numeric quantities, and long CJK strings), and layers a manual anchor checklist (anchors.md) plus the premise_anchor tool on top. When a critical premise is missing it injects a Chinese warning that points back to the log; tagged anchors carry a higher-priority prefix.
Built for writers who draft novels, world-building notes, or long non-fiction in Chinese through DSH. The upstream original only recognises ASCII anchors and is essentially idle on Chinese content; this fork shifts extraction to Chinese-first and adds the manual checklist, tool calls, and the /anchor command, turning no lost premises from a matter of luck into something checkable.
Use Cases
- Multi-chapter novel sessions where compaction silently drops key backstory premises
- Worldbuilding spans dozens of turns and the summary loses critical lore anchors
- Long non-fiction drafting where a compacted summary contradicts earlier established facts without the model noticing
Best For
- Writers crafting Chinese novels or long-form worldbuilding through DSH
- Authors maintaining character sheets and lore cards across lengthy sessions
- Users sensitive to information loss during context compaction who want an alert mechanism
Related Plugins
Memory layer for coding agents that indexes local session history and auto-recalls relevant context before edits, commands, and failures, with no manual search needed.
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.