Plugins

Git Bash shell tool for Windows: runs commands through the bash bundled with Git for Windows, with timeout, sandbox, output truncation, and background jobs, replacing the slow pwsh tool and the WSL-only bash tool.

DeepSeek Harness plugin that aggressively saves tokens without affecting task quality: lossless head/tail truncation of tool results (default), grep-result folding, pressure-triggered dsh-compaction-basic delegation, and an LLM-summary CompactionEngine subclass for the aggressive mode. Saves 97.6% of input tokens on an 8-scenario synthetic suite, with all reductions routed through dsh event flow so model-visible = logged stays intact.

Host-side continuation for turns that end without declaring completion: a closure rule (present, or a final (TASK_DONE) / (TASK_BLOCKED: reason) line) is injected into the system prompt, and a follow-up user message is sent on silent completion, retryable errors, truncation and restart-interrupted turns, with grace, cooldown and backoff; on cold start it rescans stored sessions and re-arms active goals the host disarmed. Main switches and timing live in the plugin settings card.

Codex-style `@file` mentions for the web GUI: search and reference workspace files of every format (.java, .vue, PDF, images, extension-less, hidden) with no index truncation. Replaces dsh-at-file.

Auto-continue for the dsh web GUI: when a local model hits its output-token cap and the max-tokens notice appears, the plugin automatically queues a short continue message so the reply keeps flowing — dual-source truncation detection, loop protection, and no UI.

Renders the built-in cache-hit percentage with three decimals and widens the composer stats line to prevent truncation.

Automatic recovery when a turn ends badly: resume after a transient failure, continue after output truncation, re-run after an unfinished todo list, plus a request-level 429 backstop and a client dock with edit-resend, manual retry and stop-and-re-ask.

When a turn is truncated by max-tokens, keeps the model writing inside the same turn instead of ending it with the warning bar and waiting for a manual continue; after a per-turn limit it hands over to cross-turn auto-continue plugins.

Keeps a turn from closing on an unfinished reply: a stream guard turns silent truncation into an in-step retry, and a completion protocol asks the model to declare how each turn ended.

Idle-time context truncation for a single DeepSeek Harness session. After a configurable number of idle hours the older conversation is dropped without a summary, so one long-lived session cannot leak one task into the next; it ships a per-session toggle, a tab next to the chat, and a /truncate command that IM bridges can drive.

Injects your existing markdown documents (notes, rules, knowledge-base files) into every dsh session's system prompt as read-only memory — list file or directory paths in `files`; no database, no service, no write path; UTF-8 byte-budgeted truncation with line-boundary fallback, total budget, YAML front-matter stripping, and freshness timestamps.

Long-term memory for DeepSeek Harness, read-only as of 0.1.0: a scribe_recall tool that reads a plain-markdown memory room, path rules that refuse directory traversal, Unicode smuggling and NTFS alternate data streams, and a bounded index whose truncation is always reported rather than silent.

A @-file reference tool for DSH: registers an at_file_read tool that injects workspace-file contents into the session, with traversal protection and a 50KB truncation cap (MIT).

dsh-shell for DeepSeek Harness: runs cmd / PowerShell commands on Windows (and bash-like shells on POSIX) with a bash-inspired safety model — timeouts, output truncation, exit codes, background jobs and a cwd allowlist