wangxing-git/dsh-autogate

README excerpt

Languages: English · 简体中文 dsh-autogate DeepSeek Harness auto-approval plugin: adds two permission presets — semi-auto ( auto-ask ) + full-auto ( auto ) — on top of the workspace-write sandbox , using layered decisions: deterministic rules + LLM safety approval + (in semi-auto) proactive human approval initiated by the denied party. It keeps the workspace sandbox boundary and never relaxes to full-access. Layered design Layer Decision Description L0 Deterministic rules allow / deny Zero cost, zero LLM: read-only ops, session state, in-workspace edits and deletes, build/test, and the run code container pass directly; reads of ordinary paths outside the workspace pass directly; writes/deletes outside the workspace (except writes to sensitive shell/credential config files) pass and rely on the workspace-write sandbox to intercept + escalation popup; writes to sensitive config files outside the workspace go to LLM review; empty commands, dynamic command names, and missing arguments fall through to the sandbox; privilege escalation, self-destruction, credential exfiltration, and deletion of critical paths are hard-denied L1 LLM safety approval allow / deny Operations the sandbox doesn't …

Category