TT-Wang/dsh-slice-agent-loop
C3A drop-in DeepSeek Harness agent loop whose context engine is a bounded slice instead of a growing transcript
★ 1+ · TT-Wang/dsh-slice-agent-loop source on GitHub · this plugin in the registry
TT-Wang/dsh-slice-agent-loop is a DeepSeek Harness plugin rated C3 — powerful capability combined with sensitive behavior. It patches the dsh runtime, can rewrite the system prompt, executes system commands.
What it can do
| Capability | Flag | Evidence |
|---|---|---|
| patches the dsh runtime | runtime_patch | ./cordis.patch.yml |
| can rewrite the system prompt | prompt_surface | systemPrompt |
| executes system commands | exec | ×3 src, e.g. scripts/gen-goldens.mjs:18, scripts/gen-goldens.mjs:18 |
Services it injects
agents invariants llm sessions systemPrompt tools
Hooks it attaches
agent/created agent/disposed agent/error agent/pre-step agent/request agent/request-error agent/session-start agent/status agent/turn-stopping llm/stream session/created session/disposed session/event tools/result
Environment variables it reads
DSH_HOME SLICEAGENT_REPO SLICEAGENT_PYTHON DSH_SOURCE
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.