Destructive-action reliability layer for AI agents on DeepSeek Harness: intercepts Bash before execution, quarantines deletions to .agent-trash/ with a restorable manifest, snapshots git before overwrites, and escalates compound shapes to the human (Decision Protocol: ALLOW / RELOCATE / SNAPSHOT / ASK / BLOCK). Rules live in one Python core shared across harnesses.
Agent 連携 · 開発ツール
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:mokuyoaxis/agent-guard
Agent-mode floating window for DeepSeek Harness (dsh): switch a session's system prompt between the harness's own prompt and 13 archived coding-agent CLI prompts (Claude Code, Codex CLI, Antigravity, Grok, Kimi, MiniMax, opencode, Pi…) from a draggable Web-UI panel. Cordis bundle, log-safe via the official /agent command.
Agent 連携 · インターフェース · 開発ツール · 生産性ツール
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:deepseekharness-dsh/agent-mode-change
DSH plugin: cross-session call center for agents. session_message delivers messages between DSH sessions (idle: visible full-text wake; busy: mid-turn-safe notice; plus runtime-context notes) and contacts is a named directory over session ids — resolve/call/manage contacts by alias, self-register by name only. Formerly taskboard-flow (board triggers/triage/task engine removed in v1.0.0). Config is file-based (cordis composition) — no web UI.
AGPA (Agent Player Achievements) for DeepSeek Harness — native achievement_* tools bridging the AGPA engine, plus automatic event tracking from dsh session events
Agent 連携 · 生産性ツール
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:eiainano/agpa-dsh-plugin
DSH arch-doc skill plugin: analyze a codebase and generate architecture documentation (module responsibilities, dependencies, entry points and run methods).
スキル · Agent 連携 · 開発ツール
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:duyanta123/arch-doc
Stop the drip of piecemeal questions: collect open questions as they arise into a parked list, group them by decision context, ask once per group with defaults proposed and an explicit answerable format, and keep working on everything that does not block — so the user answers in one sitting instead of being interrupted all day.
スキル · Agent 連携 · 開発ツール
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:ChenneyZhuang/ask-batch
Triage a scattered backlog into a bounded queue of decisions: every item gets a disposition (now / next / scheduled / waiting / reference / drop) with a stated reason, dropped items leave the list instead of haunting it, and anything that changes a commitment goes out as one approval batch.
スキル · Agent 連携 · 開発ツール · 生産性ツール
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:ChenneyZhuang/backlog-triage
DSH Web GUI brand customization: configure the sidebar brand mark/name, the new-session hero logo & headline, the running-turn thinking label, and the browser tab title from a General Settings row.
Agent 連携 · インターフェース · 生産性ツール
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:WayJ/dsh-work#fc381507d1a812baf4387d824c6aaf3962941cb3&path:plugins/brand-profile
Write changelog entries while the change is in hand: user-facing entries in who/what/how-to-upgrade shape, actions separated from fixes, breaking changes flagged at the top, and a verify pass against the actual diff so the log never claims what the code does not do.
スキル · Agent 連携 · 開発ツール
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:ChenneyZhuang/changelog-capture
Harness-neutral Traditional Chinese enforcer + offline converter (skill, DSH bundle, CLI). One script axis either way + Cantonese/Japanese-only filters. Converts both ways; optional wording preference. Also: an LLM output guard, cp950/GBK console scan.
スキル · Agent 連携 · 開発ツール
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:KSF1216/chinese-script-policy