55 results
DSH plugin: fixes compaction failure on local qwen3.8-27b gateways — xhigh thinking burns the entire output token budget, so thinking is off for compaction-only, with the model's non-thinking sampling parameters
Cost governance & budget enforcement for DeepSeek Harness: multi-provider pricing, per-model cost accounting, budget quotas with soft/hard thresholds, and cross-session trends.
Two-phase session captioning for the DeepSeek Harness: instant keyword captions while the conversation is busy, budget-model refinement when it goes idle.
Shared token budgets for DeepSeek Harness agent trees
Hard-stop DeepSeek Harness tool use after a per-session call budget.
DeepSeek Harness 模型额度与 Token 统计面板：只显示 DeepSeek 账户余额，并按日汇总所有 Provider 与模型的 Token 用量。
Real DeepSeek balance and spend tracker for DeepSeek Harness: official account balance and consumption in the session header, monetary spend alert with auto-stop, token breakdown by peak/off-peak pricing. Fork of ds-budget-meter.
DeepSeek Harness plugin: LLM cost metering with budget enforcement (deny/warn), peak-valley pricing and a live usage dashboard
DSH (DeepSeek Harness) Cordis plugin: retry ANY model-request failure (any error code) with exponential backoff — retry budget adjustable from Settings → General via a slider (0–64, default 8).
Independent same-host session-to-session interconnection for the DeepSeek Harness. In-process broker delivery, name (title) addressing, dormant-session wake, durable mailbox, anti-loop budget. Inspired by @dpskh/a2a injection idiom; no hub or network mesh.
Manual budget-targeted context compression for DeepSeek Harness: the conversation model picks ranges, a cheap flash-tier route writes the checkpoint summaries.
Memory Standard Protocol (mm) for the DeepSeek Harness: layered MEMORY.md + deterministic budgets + frozen snapshots + cross-agent markdown/URI interop + session-log digestion. Ships as a dsh bundle (tools: mem_read, mem_write, mem_search, mem_budget, mem_digest) and a standalone CLI.
预算分配
BSP coding swarm plugin for DeepSeek Harness: nested 3-tier swarms (root + squad leads on deepseek-v4-pro, up to 100 concurrent workers on deepseek-v4-flash) with barrier-synchronized phases, worktree isolation, guardrail grants, budget caps, and cache-stable prefixes
婚礼预算分配
保费预算
Keeps a local model's context at a size the GPU handles well: measured prefill speed, hard ceiling, early compaction for DeepSeek Harness
DeepSeek Harness guardrail that limits how many files an AI coding agent can edit per turn
Spend gate for DeepSeek Harness: refuses over-budget model calls on the llm/stream waterfall, before they are billed. Auto-downgrade past a soft limit.
Cross-session knowledge for DeepSeek Harness: ctx.knowledge service + knowledge_* tools, SQLite FTS5 trigram retrieval (CJK substring) with L1 query expansion. V1.11 contract (write/update/search/list/delete + budget + events + error codes).
Native Context Graph and budget-aware context injection plugin for DeepSeek Harness
Turn a paper into an academic talk for DeepSeek Harness (dsh) — 论文转 PPT / 组会汇报 / 会议报告 / 答辩 / 开题. Reads the real full text, budgets the slot per venue, and builds the deck around the figures with the speaker notes written as the script. Drives dsh-ai4scholar and dsh-slides.
送礼预算
DSH rc.8 Web settings row for active providers' model-request retry budget