ai-safety

Second-model AI auto-review for DeepSeek Harness approval requests: a read-only reviewer subagent returns structured allow/deny verdicts with reasons, fail-closed by default, fully auditable from the session log (approval/asked -> autoReview/verdict -> approval/decided).

Claude Code-style declarative permission rules for DeepSeek Harness: ordered allow/deny/ask rules with tool-name, argument (glob/regex), and workspace-path matching on the tools/pre-execute waterfall, session-log audit, and HMR reload.

Execution-time drift firewall for long-running DeepSeek Harness agents. Real-Harness tests: unsafe stale mutations 12/12 native -> 0/12; valid controls 7/7 both; post-SIGKILL unsafe continuation 2/2 -> 0/2.

让 DSH Coding 流程更具有确定性！以 DSH 预设与全局插件（任意会话皆可使用）两种形式交付。收敛 → 规范化 → 实现：并行 TDD 双环收敛门禁、对抗式跨源评审、对源逐条引用核查、先行技术碰撞检索——每道门控均提供诚实记录，绝不静默转绿

Deterministic fail-closed tool-call authorization for DSH with evidence: allow/block/ask policy gate plus approval-chain deferral.

DSH 远程终端插件:在 Web UI 直连 SSH,危险命令三级护栏,AI 命令须真人确认,密码加密存储、不进 AI 会话 | Remote SSH terminal plugin for DeepSeek Harness: 3-level command guard, human-approved AI commands, encrypted credentials

DeepSeek Harness 开发锻造工坊：审批守卫、开发 Skills、GitHub/浏览器能力与 Token Watch 消耗监督，装上就能干活。

Safety harness plugin for DeepSeek Harness (DSH): execution-time guard, trash-based safe_delete, composition snapshots & rollback, pre-restart validation. Standalone CLI, zero dependencies.