All Plugins

Telegram duty gateway: phone-message task loop on a dedicated duty session, global approval forwarding with inline approve/reject buttons, telegram_ask choice questions, duty/local toggle with a web duty banner, zh/en messages, zero idle tokens.

Official-grade GitHub CI integration: a composite action.yml, a polling PR review bot with idempotent inline comments and a status-check gate, plus PR/issues tools with every write gated by human approval.

Runtime-switchable model output styles with Claude Code outputStyles parity, plus the output.render.* presentation protocol: a /style command, per-session persistence, systemPrompt injection, six built-in styles, a web picker, and a renderer registry with per-session/per-tool rules and /export.

Four-module session companion: smart conversation export (Markdown/PDF/JSON/PNG long-image with privacy redaction and batch ZIP), context handoff summaries with template save/import, API cost optimization (live official pricing, peak/off-peak scheduling, daily/monthly budgets, model routing), and global conversation search with in-chat find (Ctrl+F, CSS Custom Highlight API).

Usage tiers and shareable stats cards: 🐟→🐳 ranks by monthly token burn with a locally-computed estimated percentile, precise pricing for 46 models across 6 vendors (including size-tiered Chinese pricing), backfill of pre-install sessions, and old-vs-new comparison across the 2026-08-17 rate change. Local-only, no telemetry.

Edit user and built-in system-prompt sections with live preview.

OpenCode Go subscription usage (rolling/weekly/monthly windows with reset countdowns) in the composer dock, with a built-in credential editor.

A starter template for dsh plugins covering config, tools, events, services, hooks, browser UI slots, and slash commands.

Per-subagent LLM provider/model selection with role templates.

Docker tools for agents: ps/images/logs/inspect/exec/manage with JSON output, env-configurable CLI path, approval-gated exec and zero runtime dependencies.

The model switch for DSH: point it at any OpenAI-compatible endpoint, with curated free/cheap DeepSeek provider presets and automatic fallback when a free tier rate-limits you.

Per-model token cost tracking with configurable cache-hit/miss, output and peak-window pricing, a live session cost bar, and unconfigured-model flags.

Evidence-bound negative-knowledge ledger: persists disproven paths (command_failed, file_missing) with their outcome and precondition evidence, then warns or blocks repeat attempts until the evidence changes.

Activity heatmap in the DSH Web sidebar: GitHub-style grid of daily commits, token usage, and estimated spend, with a today stats line for all-session token totals, cache hit rate, and per-model auto-priced cost.

jinguanghai/deepseek-harness-forge-plugins#evidence-first: Evidence-first guard: warns the agent when it claims success without a matching tool-execution record in recent context. jinguanghai/deepseek-harness-forge-plugins#forge-gates: Real-compute verification gates: math simplification, logic proofs, regex validation, E-prover FOL, state-machine checks, and code repair, backed by Go-compiled binaries with prebuilt Windows executables. jinguanghai/deepseek-harness-forge-plugins#forge-memory: BM25 keyword-based memory recall. jinguanghai/deepseek-harness-forge-plugins#forge-tcm: Traditional Chinese Medicine toolkit: eight-axes diagnosis and herb-pair lookup.

Fully-local image understanding & OCR via macOS Vision Framework: `ocr_image` (text, table layout + coordinates) and `view_image` (scene, faces, QR) — paste multiple images into the web input box or pass path/URL/base64; images never leave your Mac.

Native Search1API (s1) web research tools: search, news, page crawling, sitemap discovery, and trending topics as first-class `s1_*` tools, with a bundled s1 skill.

DingTalk group-robot notifications (webhook + HMAC signing, zero runtime dependencies).

RSS/Atom subscription management and parsing (rss_list/add/remove/fetch/check), persisted to settings.

Advertise image paste on text-only DeepSeek routes, describe attachments with a vision sidecar, and leave native vision models untouched.