Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed visual-direction workflow with local validation and image-helper scripts, with no hidden persistence, credential use, remote payloads, or destructive behavior found.
Use this skill for visual direction and image artifact workflows. Be aware that its helper scripts can read project files, compute hashes and metadata, and write output artifacts such as masks, manifests, review packets, and release archives, so point them only at files you intend to include in that workflow.
The self_check.py description includes compiling scripts, running tests, validating repository artifacts, and producing CI-style health output. In an agent environment, that kind of maintenance automation can trigger execution of local code and broad repository reads, creating a path to arbitrary code execution or exposure of sensitive files if used on untrusted content.
The self_check.py description includes compiling scripts, running tests, validating repository artifacts, and producing CI-style health output. In an agent environment, that kind of maintenance automation can trigger execution of local code and broad repository reads, creating a path to arbitrary code execution or exposure of sensitive files if used on untrusted content.
The self_check.py description includes compiling scripts, running tests, validating repository artifacts, and producing CI-style health output. In an agent environment, that kind of maintenance automation can trigger execution of local code and broad repository reads, creating a path to arbitrary code execution or exposure of sensitive files if used on untrusted content.
The self_check.py description includes compiling scripts, running tests, validating repository artifacts, and producing CI-style health output. In an agent environment, that kind of maintenance automation can trigger execution of local code and broad repository reads, creating a path to arbitrary code execution or exposure of sensitive files if used on untrusted content.
The self_check.py description includes compiling scripts, running tests, validating repository artifacts, and producing CI-style health output. In an agent environment, that kind of maintenance automation can trigger execution of local code and broad repository reads, creating a path to arbitrary code execution or exposure of sensitive files if used on untrusted content.
The self_check.py description includes compiling scripts, running tests, validating repository artifacts, and producing CI-style health output. In an agent environment, that kind of maintenance automation can trigger execution of local code and broad repository reads, creating a path to arbitrary code execution or exposure of sensitive files if used on untrusted content.
The self_check.py description includes compiling scripts, running tests, validating repository artifacts, and producing CI-style health output. In an agent environment, that kind of maintenance automation can trigger execution of local code and broad repository reads, creating a path to arbitrary code execution or exposure of sensitive files if used on untrusted content.
The self_check.py description includes compiling scripts, running tests, validating repository artifacts, and producing CI-style health output. In an agent environment, that kind of maintenance automation can trigger execution of local code and broad repository reads, creating a path to arbitrary code execution or exposure of sensitive files if used on untrusted content.
The self_check.py description includes compiling scripts, running tests, validating repository artifacts, and producing CI-style health output. In an agent environment, that kind of maintenance automation can trigger execution of local code and broad repository reads, creating a path to arbitrary code execution or exposure of sensitive files if used on untrusted content.
The self_check.py description includes compiling scripts, running tests, validating repository artifacts, and producing CI-style health output. In an agent environment, that kind of maintenance automation can trigger execution of local code and broad repository reads, creating a path to arbitrary code execution or exposure of sensitive files if used on untrusted content.
The self_check.py description includes compiling scripts, running tests, validating repository artifacts, and producing CI-style health output. In an agent environment, that kind of maintenance automation can trigger execution of local code and broad repository reads, creating a path to arbitrary code execution or exposure of sensitive files if used on untrusted content.
The self_check.py description includes compiling scripts, running tests, validating repository artifacts, and producing CI-style health output. In an agent environment, that kind of maintenance automation can trigger execution of local code and broad repository reads, creating a path to arbitrary code execution or exposure of sensitive files if used on untrusted content.
The self_check.py description includes compiling scripts, running tests, validating repository artifacts, and producing CI-style health output. In an agent environment, that kind of maintenance automation can trigger execution of local code and broad repository reads, creating a path to arbitrary code execution or exposure of sensitive files if used on untrusted content.
The self_check.py description includes compiling scripts, running tests, validating repository artifacts, and producing CI-style health output. In an agent environment, that kind of maintenance automation can trigger execution of local code and broad repository reads, creating a path to arbitrary code execution or exposure of sensitive files if used on untrusted content.
The self_check.py description includes compiling scripts, running tests, validating repository artifacts, and producing CI-style health output. In an agent environment, that kind of maintenance automation can trigger execution of local code and broad repository reads, creating a path to arbitrary code execution or exposure of sensitive files if used on untrusted content.
The self_check.py description includes compiling scripts, running tests, validating repository artifacts, and producing CI-style health output. In an agent environment, that kind of maintenance automation can trigger execution of local code and broad repository reads, creating a path to arbitrary code execution or exposure of sensitive files if used on untrusted content.
The self_check.py description includes compiling scripts, running tests, validating repository artifacts, and producing CI-style health output. In an agent environment, that kind of maintenance automation can trigger execution of local code and broad repository reads, creating a path to arbitrary code execution or exposure of sensitive files if used on untrusted content.
The self_check.py description includes compiling scripts, running tests, validating repository artifacts, and producing CI-style health output. In an agent environment, that kind of maintenance automation can trigger execution of local code and broad repository reads, creating a path to arbitrary code execution or exposure of sensitive files if used on untrusted content.
The skill explicitly routes to scripts and helpers that imply shell, filesystem, and potentially network-capable execution, yet it declares no permissions or allowed-tools scope. That creates an overly broad execution surface where an agent may invoke powerful tools without an explicit least-privilege contract, increasing the chance of unintended file access, command execution, or external data exposure.
The markdown mandates a specific labeled output format using Chinese text (生成前设计说明) and later provides additional required Chinese section labels and response structure. This imposes a language choice on the interaction without offering the user any locale or language option, which fits the language/locale policy violation category.
The Communicate compactly section defines the required direction and generative-attempt templates entirely with Chinese headings and labels, with no opt-in or alternative locale path. Because this is a natural-language instruction in the skill definition, it constitutes a forced language constraint.
The phrase 'turn the required slogan into restrained bilingual editorial typography' imposes a language presentation choice in natural-language guidance. This can violate language/locale policy because it does not indicate user opt-in or explain that bilingual output is required for a specific regional or compliance reason.
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
- “你没审美”, “土”, “廉价的人才会喜欢”, or any attack on identity or competence.
- “高级”, “电影感”, “年轻人喜欢”, “色彩心理证明” without a defined mechanism and context.
- “完全保持”, “像素未变”, or “已经修复” without verification.
Prefer:
subprocess module calls execute external commands. Without careful input validation, this enables command injection.
for path in sorted(ROOT.rglob("*.json")):
load_json(path, blockers)
tests = subprocess.run(
[sys.executable, str(ROOT / "scripts" / "run_tests.py")],
text=True,
capture_output=True,
subprocess module calls execute external commands. Without careful input validation, this enables command injection.
def run_script(name: str, *args: Path | str) -> subprocess.CompletedProcess[str]:
return subprocess.run(
[sys.executable, str(SCRIPTS / name), *(str(arg) for arg in args)],
text=True,
capture_output=True,
No suspicious patterns detected.