插件

PLUR memory for DeepSeek Harness — engrams injected straight into the prompt, no tool call required

渐进披露 MCP 网关：用 `mcp_search` 检索大型远程工具目录，再由 `mcp_call` 按精确 schema 调用，并采用惰性连接与有界缓存。

为长期或定义不完整的 Harness 任务提供持久执行合同、递归工作图、关键澄清与证据门禁。

dsh 分层记忆管线：自动把会话蒸馏为原子事实、场景摘要与用户画像（L0~L3），BM25 + 向量混合检索，chat/work 双族隔离，每步模型调用前注入相关上下文。

Markdown folder long-term memory for DeepSeek Harness: one file per fact, an index line per memory, mounted into every request as a system prompt section

Read-only performance diagnostics for DeepSeek Harness: session load (open/restore) timing, spill-hit counts, compaction count and trigger, context-injection volume (AGENTS.md/skills/tool-schema token share), and LLM cache hit rate — surfaced via the /fas