dsh-brainagent
BrainAgent — brain-inspired plugin for DeepSeek Harness (dsh): a pipeline of heuristic filters, statistical memory stores and context injections; reward-ledger + UCB1 bandit learning loop (RL-lite)
86 results
BrainAgent — brain-inspired plugin for DeepSeek Harness (dsh): a pipeline of heuristic filters, statistical memory stores and context injections; reward-ledger + UCB1 bandit learning loop (RL-lite)
Local prompt-injection and secret-exfiltration firewall for DeepSeek Harness
Persistent 5-layer memory system for DeepSeek Harness (entry injection, relevance retrieval, agent tools, and LLM-backed auto-extraction).
Mobile adaptation for the DeepSeek Harness Web UI: compact typography, drawer sidebar, two-page settings, responsive tables and stats, safe touch targets. Pure CSS injection, no source modification.
GPU-aware execution layer for DeepSeek Harness: gpu_status / gpu_exec / gpu_run_bg tools, per-step GPU context injection, automatic CUDA_VISIBLE_DEVICES card selection
Source-aware prompt injection protection for DeepSeek Harness
Agent-driven long-term memory for DeepSeek Harness: scoped memory (global + per-workspace), layered entries (fact/knowledge/episodic), time-bucket compaction (day→week→month→year), associative recall (related chains) + memory_relate navigation (multi-hop BFS closure), auto recall injection on user messages (CJK bigram search, tail append), agent-decided content.
DSH WebUI capability controls for skills, MCP servers, tools, prompt injections, approvals, and safety guards at session, project, and global levels.
Yunoseek pink skin for the DeepSeek Harness web GUI (pink --dsw-* token layer, yunoseek brand artwork, assistant avatar) plus the Yunoseek system-prompt injection: the identity prompt rendered as the first system-prompt section.
Mid-turn user steering for DeepSeek Harness: /steer (visible) and /inject (quiet context notice) slash commands, a model-facing mid-turn guidance prompt section, and an always-visible injection bar above the composer.
A DeepSeek Harness plugin that integrates claude-mem: query, context injection, manual save, and session lifecycle over a local claude-mem worker's HTTP API
Mingleng mcpguard for DeepSeek Harness 鈥?the first security plugin for DSH. Scans skills and MCP configs for prompt injection, homoglyphs, hidden Unicode, dangerous shell and credential leaks.
Persistent memory for DeepSeek Harness: per-conversation notes, selective injection, project memory, semantic Vault search (bge-m3, free optional key) and automatic retrieval — agents stop forgetting.
Focus board for DeepSeek Harness agents: durable, model-maintained notes in the session workspace that pin the objective, constraints, and decisions across compaction and sessions — with automatic context injection, an archive on clear, and an optional read-only web panel
Private, per-project system-prompt rules for DeepSeek Harness — matched by git remote URL, repo path, or cwd prefix; worktree-aware; never committed to the repository.
Scan agent skill packages for static security risks. Use for skill audit, security check, prompt-injection review, suspicious scripts, unsafe dependencies, zip package review, and pre-install skill review. Not for runtime monitoring.
A safe external-skill discovery and adoption bundle for DeepSeek Harness
Claude Code-style long-term memory for dsh: markdown memory files with frontmatter, MEMORY.md index, session-start prompt injection, and a memory tool.
NoAtMark text hygiene as a DeepSeek Harness (dsh) plugin — sanitize untrusted text, scan invisible characters, clean LLM formatting, and escape CSV formula injection.
Thread session memory for dsh — lossless event capture + status-card injection (dsh plugin, standalone repo).
Task-to-skill pairing with a laziness ladder and a safe install cellar. Never overpours, never serves untasted. DSH · Claude Code · Codex
Eight DeepSeek Harness plugins: persona, language guard, vision fallback, python workdir guard, windows encoding guard, cross-agent memory, image generation, and skill shell injection.
Portable DSH WebUI skill security inspection with local rules and optional model review
Persistent cross-session memory plugin for DeepSeek Harness: workspace-scoped durable facts with model-facing memory_write/list/search/forget tools and bounded prompt injection