201 results
Inject Context for DeepSeek Harness (DSH): persist your own context entries and inject them as a standalone 'Remember:' row — right after your message — per turn or once per session (per-entry 'Every turn' switch). Simple/Advanced datasets, native Settings editor, standalone manager page.
Indirect prompt-injection guard for DeepSeek Harness: taints tool output by origin and gates privileged tool calls that follow untrusted content
Layered, approval-gated, auditable cross-session memory for DeepSeek Harness — a capability seam (ctx.memory service + local SQLite provider + memory tool + frozen snapshot injection), not another memory warehouse
DSH plugin: inject dynamic/random HTTP request headers (${uuid} templates) into LLM provider requests, e.g. a run-stable x-opencode-session for the opencode-go gateway.
Add mood-aware Japanese kaomoji to DeepSeek Harness (dsh) replies; prompt-injection plugin with a curated library from kaomojiya.org.
Calendar and clock for DeepSeek Harness: set an alarm that wakes a chosen conversation at a chosen instant, by keyword, with the system time and the drift injected into the wake. In-process timer plus a Windows Task Scheduler mirror.
User-only instruction memory for DSH: long-term instructions maintained in the settings UI and auto-injected into every conversation — the model has no write access. 指令记忆：只由用户在设置页维护的长期指令，自动注入此后每轮对话；模型没有任何写入口。
RAG-style skill router for DeepSeek Harness: embeds skills, matches user tasks by similarity, and auto-injects the relevant skill bodies.
Static trust audit for DeepSeek Harness plugins: capabilities, injected prompts, token cost, provenance, and a reproducible 0-100 trust score