strukto-ai/mirage

C3

The World's First Virtual Terminal for AI Agents

★ 3.7k · strukto-ai/mirage source on GitHub · this plugin in the registry

strukto-ai/mirage is a DeepSeek Harness plugin rated C3 — powerful capability combined with sensitive behavior. It patches the dsh runtime, executes system commands, decodes base64 payloads, starts a network server.

Installable plugin — declares a dsh.bundle manifest

What it can do

CapabilityFlagEvidence
patches the dsh runtimeruntime_patch./cordis.patch.yml
executes system commandsexec×50 in authored code, e.g. integ/ntn_conformance.ts:24, integ/ntn_conformance.ts:24
decodes base64 payloadsbase64_decode×12 in authored code, e.g. integ/server/github/contents.ts:260, integ/server/github/pulls.ts:978
starts a network servernet_server×7 in authored code, e.g. integ/server/github/search_conformance.ts:89, integ/server/hf_hub/mcp.ts:1617
reads credential-class env varstoken_envAWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, E2B_API_KEY, HF_TOKEN, LANGFUSE_PUBLIC_KEY, LANGFUSE_SECRET_KEY

Outbound domains

api.github.com docs.github.com huggingface.co hf.co avatars.githubusercontent.com www.yamllint.com docs.google.com www.notion.so airtable.com opengraph.githubassets.com

Environment variables it reads

S3_ENDPOINT AWS_ACCESS_KEY_ID AWS_SECRET_ACCESS_KEY REDIS_URL S3_REGION MONGODB_URI HTTP_ENDPOINT CHROME_PATH S3_BUCKET NTN_BIN

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.