Borrowing it
Nothing to install: this file belongs to JinNing6/CyberHuaTuo-Plugin. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/JinNing6/CyberHuaTuo-Plugin/main/.agents/skills/cyberhuatuo-agent-guard/SKILL.mdgit clone --depth 1 https://github.com/JinNing6/CyberHuaTuo-PluginWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/jinning6/cyberhuatuo-plugin/cyberhuatuo-agent-guard)<a href="https://agentmods.dev/skills/jinning6/cyberhuatuo-plugin/cyberhuatuo-agent-guard"><img src="https://agentmods.dev/badge/skills/jinning6/cyberhuatuo-plugin/cyberhuatuo-agent-guard/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/jinning6/cyberhuatuo-plugin/cyberhuatuo-agent-guard"><img src="https://agentmods.dev/badge/skills/jinning6/cyberhuatuo-plugin/cyberhuatuo-agent-guard.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00051 | $0.00747 |
| Opus 5 | $0.00026 | $0.00374 |
| Sonnet 5 | $0.00010 | $0.00149 |
| Haiku 4.5 | $0.00005 | $0.00075 |
Grade C, and why
cyberhuatuo-agent-guard scanned grade C with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
- Permanent or recursive deletion such as `rm -rf`, `Remove-Item -Recurse`, `del /s`, `find -delete`, or shredding. How it starts
The opening of the file, as written. The whole thing — 73 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CyberHuaTuo Agent Action Guard
Trigger
Run this skill before any proposed action that can destroy or broadly remove user data, including:
- Permanent or recursive deletion such as
rm -rf,Remove-Item -Recurse,del /s,find -delete, or shredding. - Destructive Git operations such as
git clean,git reset --hard, checkout/restore that discards work. - Disk formatting, partitioning, wiping, raw-device writes, or free-space erasure.
- Database
DROPorTRUNCATEstatements. - Mirror synchronization that deletes destination files, such as
robocopy /MIRorrsync --delete.
Preflight
For first-run verification, review the built-in examples without executing anything:
cyberhuatuo guard --self-test --workspace-root "<workspace>"
Continue only when it prints SELF-TEST PASSED with one ALLOW, one ASK, and one BLOCK.
Prefer the MCP tool when available:
agent_action_guard(
command="<exact proposed shell command>",
cwd="<current working directory>",
workspace_root="<declared workspace root>"
)
Otherwise use the CLI:
cyberhuatuo guard "<exact proposed shell command>" --cwd "<cwd>" --workspace-root "<workspace>" --json
The review is read-only and does not execute, rewrite, approve, or intercept the command.
Case Report
When a real decision differs from the expected result, generate a local redacted case without executing the reviewed command:
cyberhuatuo guard "<exact proposed shell command>" --cwd "<cwd>" --workspace-root "<workspace>" --expected BLOCK --report reports/guard-report.md
Review the complete preview before confirming the write. Repeat --redact <literal> for private project or business identifiers. The report performs no network request or automatic upload, refuses implicit overwrite, and fails closed when private-key material or a known unredacted secret remains.
Use public Guard false-positive, ordinary false-negative, or integration-gap forms only for safely redacted details. Route reliable Hook, parser, wrapper, protocol, or enforcement bypasses through the repository's private vulnerability reporting page.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 11d ago First seen · 73 lines · 51 tokens per session scan C ab7072726251
cyberhuatuo-agent-guard is a skill published in the GitHub repository JinNing6/CyberHuaTuo-Plugin (3 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 51 tokens to every session and 747 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
ziran
ZIRAN is an open-source security testing framework for AI agents. It discovers dangerous tool chain compositions via knowledge graph analysis, detects execution-level side effects (not just text output), and runs multi-phase trust exploitation campaigns that model real attacker behavior.
sidclaw-governance
Add policy evaluation, human approval, and audit trails to any tool. Powered by SidClaw.
docs
ACMI is a universal architectural framework for giving AI agents persistent, real-time context. It replaces fragmented, multi-table database joins with a single, lightning-fast Key-Value engine (Upstash Redis) optimized specifically for LLM context windows.
tokendiet
Audit this codebase for wasteful LLM API spend and hand back a concrete, honest fix plan — missing prompt caching, uncapped retries, prompt bloat, no batching, overpowered models. Describes technical waste and published provider rates only; never fabricates a dollar figure. Trigger on requests like "audit my LLM…
aura
Give this agent a real email address, a webhook URL, durable memory across runs, and the ability to wait for an event without burning tokens. Use when a task needs a verification code sent by email, needs to wait for a webhook or an approval, needs to remember something after the session ends, or needs to survive a…
boundedrelay-adaptive-sdd
Coordinate the optional BoundedRelay Spec Kit workflow with strict independent reviews, deterministic effort routing, bounded proposals, convergence, and handoff.