Sev7eEn7/dsh-sieve

C3

dsh-sieve: context engineering & token optimization plugin for DeepSeek Harness (DSH) — tool output filtering, context pruning, progressive skill disclosure. 36% smaller payload in offline replay. DSH

★ 100+ · Sev7eEn7/dsh-sieve source on GitHub · this plugin in the registry

Sev7eEn7/dsh-sieve is a DeepSeek Harness plugin rated C3 — powerful capability combined with sensitive behavior. It patches the dsh runtime, can rewrite the system prompt, gates tool execution, executes system commands.

Installable plugin — declares a dsh.bundle manifest

What it can do

CapabilityFlagEvidence
patches the dsh runtimeruntime_patch./cordis.patch.yml
can rewrite the system promptprompt_surfaceapproval, systemPrompt, tools/pre-execute
gates tool executiontool_gatehook: tools/pre-execute
executes system commandsexec×5 in authored code, e.g. scripts/release-pack.mjs:11, scripts/release-pack.mjs:11

Tool registrations

1 confirmed registration point(s); analysis is incomplete.

Services it injects

agentTeams agents approval commands connection llm locale permissionPresets sessionProjections sessions sidebarRightTabs sieve skills slots spillStore storageDomain subagents systemPrompt tokenMeter tools

Hooks it attaches

agent/assistant-stream agent/created agent/disposed agent/pre-step agent/request agent/status agent/turn-stopping connection/reset session/event sieve/trouble tools/post-execute tools/pre-execute tools/result

Outbound domains

openrouter.ai api.typesafe.ai

Environment variables it reads

ALLOW_UNSUPPORTED_DSH

How to read this

Levels measure capability surface and transparency, not maliciousness. A C3 plugin can be entirely legitimate — a desktop shell genuinely needs subprocesses. The point is that you can see this before installing. See the levels explained and how dsh plugins work.

Findings come from static analysis of shipped code; nothing is executed. Think a flag is wrong? Open an issue — every flag cites the file and line it came from.