Back to skill

Security audit

BII 简报 × 贝莱德智库风格(pptx+pdf+html 三格式)

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent document rebranding/export workflow, with disclosed local file generation and conversion behavior but no artifact-backed malicious activity.

Before installing, confirm you are allowed to apply a BlackRock-style visual identity and review outputs so they cannot be mistaken for official BlackRock material. Treat generated PPTX files as containing hidden/searchable speaker-note text copied from the PDF, and use a fresh output directory to avoid overwriting prior exports.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (19)

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

代码行为与声明的核心用途明显不一致。该脚本只验证输入 HTML 是否存在,然后调用上游脚本执行页面尺寸/溢出检查,属于“体检”子步骤中的一个包装器。它没有任何对 HTML 内容、样式、品牌元素或导出流程的处理逻辑,也没有生成 pptx、pdf 或 html 的能力(输入本身就是 html)。虽然声明中提到“七门体检”,当前代码最多只覆盖其中逐页版式检查的一部分,远不足以支撑所宣称的主要功能,因此应判定为描述与实际代码严重不匹配。

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

声明描述的是一个完整的品牌改版与多格式交付工具,重点在内容/版式/配色的重定向与质量校验;而代码只是对“已定稿的贝莱德风格 PDF”做后处理,生成一个由整页位图组成的 PPTX,并在备注中存放提取文本。输入前提都不同:声明面向 BII 简报改版,代码假定输入已经是定稿 PDF。输出能力也明显不足:代码不生成 HTML 或新 PDF,不编辑页面内容,不进行风格转换。虽然代码提到贝莱德风格和页脚免责声明,但那只是针对已有 PDF 的备注清洗,不等同于声明中的核心能力。因此属于明显的描述与行为不匹配。

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The README instructs maintainers to export a GitHub token and run a publishing script that performs remote actions, but it does not explicitly warn that the token is sensitive or that the command will publish externally. This can lead to accidental credential exposure in shell history, logs, screenshots, or copy-paste reuse, and can cause unintended remote releases if a maintainer runs it without fully understanding the side effects.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
89% confidence
Finding

The skill documentation describes capabilities that require file reads/writes and shell execution, but it does not declare an explicit tool scope such as permissions or allowed-tools. In an agent environment, missing scope boundaries can cause the skill to run with broader ambient authority than intended, increasing the blast radius if the skill is misrouted, modified, or abused.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger phrases are broad enough to overlap with normal conversation about branding, layouts, or making another version of a document. Overbroad routing can cause this file-writing, shell-using skill to activate unexpectedly on unrelated prompts, which is risky because it may operate on local files or invoke external tooling without the user intending that specific workflow.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

This markdown file uses Chinese throughout, including the title, section headings, and operational guidance, but does not indicate that Chinese is optional or that the skill is region-specific. Under the language/locale policy, forcing a single language without user opt-in can be a natural-language policy violation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The natural-language instructions and documentation in this file are fully Chinese, and there is no indication that users may choose another language or that the skill is intentionally limited to a Chinese-language audience. Per the policy, forcing a specific language without user opt-in is a locale/language policy concern.

Content

No source excerpt is available for this finding.

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/check_br.py (reported line 141)May include surrounding context.

python
# ── 门 6 · PDF 页数与页面尺寸 ─────────────────────────────────────────
    if a.pdf and os.path.exists(a.pdf):
        info = subprocess.run(['pdfinfo', a.pdf], capture_output=True, text=True).stdout
        npg = int(re.search(r'^Pages:\s+(\d+)', info, re.M).group(1))
        m = re.search(r'^Page size:\s+([\d.]+)\s+x\s+([\d.]+)', info, re.M)
        w, h = float(m.group(1)), float(m.group(2))

Ssd 3

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The documented design intentionally preserves full extracted page text in PPT notes, which can replicate sensitive content into a new artifact with broader discoverability. In a document-conversion skill, this is more dangerous because users may distribute the PPTX assuming it only contains slide images, while the notes remain searchable and extractable.

Content

No source excerpt is available for this finding.

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/mkpptx.py (reported line 37)May include surrounding context.

python
def pdfinfo(pdf):
    out = subprocess.run(['pdfinfo', pdf], capture_output=True, text=True).stdout
    pages = int(re.search(r'^Pages:\s+(\d+)', out, re.M).group(1))
    m = re.search(r'^Page size:\s+([\d.]+)\s+x\s+([\d.]+)\s+pts', out, re.M)
    if not m:

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The script silently extracts each page's text and stores it in PowerPoint speaker notes, creating a second copy of document contents that may not be obvious to recipients. This can expose sensitive text, hidden OCR output, or material users believed was only present visually in the PDF, increasing unintended disclosure risk when the PPTX is shared.

Content

No source excerpt is available for this finding.

Ssd 3

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The implementation writes the full extracted text of each PDF page into the presentation notes field, which can leak embedded, OCR-derived, or otherwise non-obvious text to downstream recipients. Because the tool is specifically for reformatting and exporting investment reports into shareable formats, the copied notes materially increase the chance of inadvertent confidential-data disclosure.

Content

No source excerpt is available for this finding.

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/mkpptx.py (reported line 46)May include surrounding context.

python
def page_text(pdf, n, keep_footer):
    t = subprocess.run(['pdftotext', '-f', str(n), '-l', str(n), '-layout', pdf, '-'],
                       capture_output=True, text=True).stdout
    lines = [re.sub(r'\s+$', '', l) for l in t.splitlines()]
    lines = [l for l in lines if l.strip()]

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/mkpptx.py (reported line 56)May include surrounding context.

python
def render_pages(pdf, dpi, tmpdir):
    subprocess.run(['pdftoppm', '-r', str(dpi), '-png', pdf,
                    os.path.join(tmpdir, 'pg')], check=True)
    return [os.path.join(tmpdir, f)
            for f in sorted(os.listdir(tmpdir)) if f.endswith('.png')]

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The header comments and step descriptions are entirely in Chinese, and user-facing status/error messages later in the script also use Chinese only. This imposes a specific language on users without any opt-in or documented locale-specific justification, which matches the language/locale policy violation criteria.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

Natural-language policy checks apply to all file types. The user-facing instructions and descriptions are entirely in Chinese, and the file does not indicate that the skill is China-specific or provide an opt-in or alternative language, which can amount to a forced locale/language experience.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

This markdown file includes a reproduction command that runs a pipeline against a local HTML file and writes outputs to ./out, while earlier lines describe produced HTML/PDF/PPTX artifacts. The description does not warn the user that running the command will create or potentially overwrite local files, which is a user-impacting filesystem operation.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

This code performs a subprocess-style handoff via exec bash to another script, which is a safety-relevant operation for code files. In this file, the action is implemented without a confirmation prompt or explicit user-facing warning at the point of execution; the comments describe the delegation rationale but do not clearly disclose to the user that another script will be executed.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

Line L093 saves output to the destination path, which is a file write operation. The script prints a success message after writing, but there is no pre-write confirmation, overwrite safeguard, or explicit notice if an existing file may be replaced.

Content

No source excerpt is available for this finding.