mokuyoaxis/agent-guard
C3Make destructive AI-agent actions reversible by default — quarantine + audit + human escalation for rm/git destructive operations. Reliability infrastructure, not a sandbox.
★ 10+ · mokuyoaxis/agent-guard source on GitHub · this plugin in the registry
mokuyoaxis/agent-guard 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, command execution in build output only.
Installable plugin — declares a dsh.bundle manifest
What it can do
| Capability | Flag | Evidence |
|---|---|---|
| patches the dsh runtime | runtime_patch | ./adapters/dsh/cordis.patch.yml |
| can rewrite the system prompt | prompt_surface | systemPrompt, tools/pre-execute |
| gates tool execution | tool_gate | hook: tools/pre-execute |
| command execution in build output only | exec_bundled | ×1 in build output only, e.g. adapters/dsh/lib/read_result_guard.js:5 |
| starts a network server | net_server | ×1 in authored code, e.g. adapters/claude/harness/mock_anthropic_api.mjs:21 |
Services it injects
shell systemPrompt tools
Hooks it attaches
tools/post-execute tools/pre-execute
Environment variables it reads
MOCK_PORT MOCK_SCRIPT MOCK_LOG
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.