SKILLEMALL.ai
GitHub ↗ RU Sign in

BD yotta-verify

元信 —— 装任何技能/包前的确定性安全扫描器：prompt injection（提示注入）+ 危险模式 + SKILL.md 完整性 + 权限需求，输出 verdict（SAFE TO INSTALL / INSTALL WITH CAUTION / REVIEW REQUIRED / DO NOT INSTALL）+ audited 徽章。触发：安装/评估任何技能或 npm 包前、给技能做安全验证、生成 audited 徽章、CI 装前闸门；或用户说 装前扫描/验证/audited/安全验证/verify-skill/可信 等。边界：只做确定性静态扫描与报告，不执行被测代码、不联网、不装包、不修复；结论需人工确认，不代替最终决策。

ClawHub Agent Skills author: YottaMeta v0.3.0 MIT-0 17 files · 1 script body ≈ 946 tokens Open the sourceclawhub.ai↗ analyzed 18 h ago

元信 —— 装任何技能/包前的确定性安全扫描器：prompt injection（提示注入）+ 危险模式 + SKILL.md 完整性 + 权限需求，输出 verdict（SAFE TO INSTALL / INSTALL WITH CAUTION / REVIEW REQUIRED / DO NOT…

As a process D 46/100 · Unfinished process — weak spots: result and completion, when it triggers, inputs and preconditions

AnalyzerAI and agentstype and topics are labelled automatically from the skill text
Runs in: OpenClaw Claude / Claude Code ChatGPT / Codex Gemini CLI Cursor GitHub Copilot / VS Code Qwen Code Kimi Code CLI Z.ai (GLM) DeepSeek (Deep Code) Mistral Vibe LangChain (deepagents) Goose OpenCode
JSON
Technical rating
B
79/100
safety, quality, tests
Safety 60%
83
Quality 40%
74
Run on models
none yet
Process rating
D
46/100
Unfinished process
Result and completion w 14
0
Inputs and preconditions w 11
0
Failures and branches w 10
0
the three weakest of ten parameters · all ten

What is at stake

Medium-severity findings: the skill is probably honest, but read what alarmed the scanner.

Dangerous commands medium severity

Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.

If you install

The skill contains commands that delete files, rewrite disks or execute code fetched from the network. The agent may run them without asking if it believes the instructions require it.

For the author

Replace destructive commands with safe equivalents that ask for confirmation, scope them to one folder, and stop piping curl into a shell: pin a version and a checksum.

How to improve

  1. Say in the description WHEN to use the skill ("use when…", example requests): that is the agent's main cue.
For the model run — optional
  • Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
  • A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.

Guard findings · 9

✓ No critical or high findings

Medium and low: 9
  • medium Dangerous commands cmd-encoded-exec scripts/verify_rules.py:76
    Executes a base64/encoded payload (detector / deny-list definition; string literal in code, not executed)
    r"(?i)(?:(?:exec|eval|system)\s*\(\s*(?:base64\.)?b64decode\s*\(|(?:base64\.)?b64decode\s*\([^)]*\)\s*[^\n;]{0,60}\b(?:exec|eval|system)\b)",
    detectorcode literal
  • medium Dangerous commands cmd-persistence scripts/verify_rules.py:97
    Persistence mechanism (cron / launchd / scheduled task / autorun registry) (string literal in code, not executed)
    "macOS 启动代理/守护（LaunchAgents/LaunchDaemons plist）持久化", 70),
    code literal
  • low Risky intent intent-offensive-security README.md:33
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
    - **Prompt injection detection** — 8 categories / 28 rules (instruction override, role spoofing, encoded instructions, data exfiltration, delimiter escape, tool self-execution, hidden intent, credenti
  • low Risky intent intent-offensive-security README.md:34
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
    - **Shared malicious-pattern rules** — 54 rules kept in sync with yotta-security-audit (download-and-exec, obfuscation, persistence, exfiltration, credential theft, network calls, privilege escalation
  • low Dangerous commands cmd-persistence scripts/verify_rules.py:90
    Persistence mechanism (cron / launchd / scheduled task / autorun registry) (detector / deny-list definition; string literal in code, not executed)
    r"(?i)\bcron\b[^\n;]{0,40}(?:@reboot|@daily|@hourly)",
    detectorcode literal
  • low Dangerous commands cmd-privilege scripts/verify_rules.py:189
    Privilege escalation / world-writable permissions (string literal in code, not executed)
    "chmod 设置 setuid/setgid/sticky 权限位", 85),
    code literal
  • low Dangerous commands cmd-privilege scripts/verify_rules.py:192
    Privilege escalation / world-writable permissions (string literal in code, not executed)
    "chmod 777 全权限", 70),
    code literal
  • low Dangerous commands cmd-privilege scripts/verify_rules.py:195
    Privilege escalation / world-writable permissions (string literal in code, not executed)
    "调用 setuid/setgid", 80),
    code literal
  • low Dangerous commands cmd-pipe-to-shell scripts/verify_rules.py:415
    Downloads and executes remote code from an unrecognised host (pipe to shell) (detector / deny-list definition; string literal in code, not executed)
    POSTINSTALL_SUSPICIOUS = r"(?i)(curl|wget|bash|sh\s|python|node\s+-e|eval|powershell|/tmp|%temp%)"
    detectorcode literal

Files scanned: 17. Evidence is masked. Grey chips explain why severity was lowered.

Against the Agent Skills spec

  • warning description-no-when description does not say WHEN to use the skill (no "use when")

Process rating: all ten parameters 46/100

  • 0Result and completion. Does not say what the result is
  • 0Inputs and preconditions. Does not say what the process needs to start
  • 0Failures and branches. Linear process with no failure handling
  • 0Progress reporting. Says nothing while it works
  • 20When it triggers. No condition that starts the skill
  • 60Tools and files. Uses tools (python) that frontmatter does not declare
  • 100Steps. 30 steps
  • 100Consistency. Name and required fields are in place
  • 100Execution cost. Instruction body is 946 tokens
  • 100Running it twice. No mutating operations
  • low 11 top-level sections: this looks like several domains in one skill

Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.

Quality signals

  • +5Description has no quoted example phrases that should trigger the skill
  • +4Description does not say when NOT to use the skill (false activations)
  • +3Output format is not stated: the model decides each time
  • -32 of 4 scripts are never mentioned in SKILL.md
  • +2Single-language instructions
  • +3Description length 321: enough signal without eating the budget
  • +4Structure: 12 headings
  • +3Step-by-step instructions: 30 items
  • +4Has examples (1 code blocks)
  • +4Reference files are cited in the instructions (3 of 3)
  • +1License stated

Quality base 70; lint remarks subtract, signals add up to 100. Result: 74.

External checks

ClawHub: suspicious
This is a legitimate-looking local security scanner, but its scanner and installer have broad or unsafe edge cases that users should review before trusting it.
LLM: suspicious (high) · 13 Sept 2026