guardrails 17

Auto mode for DeepSeek Harness (dsh): an Auto entry in the Permissions selector whose approval prompts are answered by OpenGuardrails policy instead of a human — plus the full OGR guard engine underneath. No core changes.

Gives DeepSeek Harness three automatic Jev judgments — prune oversized tool output, screen fetched pages for injected instructions, suggest which skill to use — plus the jev_ask and jev_gate tools, a persistent judgment ledger, and a /jev-status diagnosis

DeepSeek Harness adapter for Koma Miko Agent Specs.

SpecWave — multi-host coding CLI (Cursor / Claude Code / DSH / agents) + P0 gates & Harness process track. Spec-driven discipline: mechanical gates (exit 2), S2 audit trail, hat roles. Formerly SpecGate / dsh-coding-kit. npx spec-wave

Allow, ask, or deny DeepSeek Harness tool calls before execution

Experimental native DSH profile bundle for AsYouMeant.

Commitment write-gate for AI coding agents: two-tier (deterministic + LLM judge) pre-execution policy. Engine-agnostic core with a DeepSeek Harness (dsh) adapter.

Jev (TypeSafe System One decision model) for DeepSeek Harness — seven tools: jev_decision (up to 25 typed questions in one parallel call, measured median 266-484 ms), jev_choose (rank 2-10 approaches), jev_batch (one set of up to 5 judgments over 1-20 tex

Self-protection guardrails for the DeepSeek Harness: protected-path interception, backup-before-destroy, CLI self-recovery commands, and strict sandbox defaulting

Jev (TypeSafe System One) judgments at the DeepSeek Harness agent-loop gates: pre-step, pre-execute, post-execute, and turn-stopping. No UI dependencies; mount it in any Harness composition.

Operator policy and a tamper-evident decision log for DeepSeek Harness. Every tool call judged against your organisation's pack, and written down.

The authority layer for agentic AI, as a DeepSeek Harness plugin. Governs every tool call against your business rules via the TapPass /v1/govern policy decision point.

DeepSeek Harness（DSH）对话互通插件：让同一进程里的对话互相发现、直接发消息、盯进度、设护栏——不用先绑定，也不用开子代理。Peer messaging, supervision, and tool guardrails between conversations.

Rate & budget governance for DeepSeek Harness — 429 backoff windows, daily token budgets, tool-call storm control; monitor by default, enforce opt-in, JSONL audit.

Declarative tool-call permission control for DeepSeek Harness — allow/deny/ask rules over tools, paths and commands, fail-closed by default, with JSONL audit log.

PII detection & exfiltration guard for DeepSeek Harness — scans tool arguments and results for personal data (emails, phone numbers, CN IDs, cards, keys), blocks outbound calls carrying PII, JSONL audit without storing raw matches.

DeepSeek Harness plugin: Jev-backed guardrails for prompts, tool calls, tool results, and model responses.