gezi-wen/sage-mem

C3

Cross-session memory for DeepSeek Harness — plain markdown files, no database, no worker.

★ 10+ · gezi-wen/sage-mem source on GitHub · this plugin in the registry

gezi-wen/sage-mem is a DeepSeek Harness plugin rated C3 — powerful capability combined with sensitive behavior. It patches the dsh runtime, can rewrite the system prompt, command execution in build output only, executes commands in development tooling the package does not ship.

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_surfacesystem-prompt/assemble
command execution in build output onlyexec_bundled×1 in build output only, e.g. lib/memory/open-folder.js:22
executes commands in development tooling the package does not shipexec_tooling×7 in development tooling the package does not ship, e.g. scripts/make-readme-images.mjs:19, scripts/make-readme-images.mjs:19
reads credential-class env varstoken_envGH_TOKEN, GITHUB_TOKEN

Tool registrations

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

Services it injects

remote slots

Hooks it attaches

fs/edit-intent fs/write-intent system-prompt/assemble tools/result

Outbound domains

api.github.com

Environment variables it reads

DSH_HOME SAGE_MEM_DEBUG SAGE_MEM_DIR GITHUB_TOKEN GH_TOKEN SAGE_MEM_STATE_DIR SAGE_MEM_VALIDITY SAGE_MEM_TASK_ID

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.