Back to skill

Security audit

blog-manager

Security checks for vulnerabilities and agentic risk

Overview

This appears to be a real blog-management CLI, but it can create, upload, and delete live blog content without authentication, confirmations, or secure transport requirements.

Use this only against a blog instance you control, preferably a test or private admin endpoint. Prefer HTTPS, avoid passing real passwords with --pwd, confirm every delete or upload target yourself, and pin dependencies before production use.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (2)

T09 · Insecure Skill Coding Practices

Warning
Location
main.py:92
Finding

Passwords May Be Exposed Through Command-Line Arguments and Plaintext HTTP

Content
View full analysis
Tuple[Any, str]: payload = { "uname": uname, "phone": phone, "pwd": pwd, "email": email, "img": img, } return client.post(USERS_PATH, json=payload), "id_response" ``` ```python # blog_manager/client.py:68-78 raw = (base_url if base_url else os.environ.get(BASE_URL_ENV, "")).strip() url = raw.rstrip("/") if not url: raise BlogConfigError( f"{BASE_URL_ENV} is not set. " f"Export it first, e.g. export {BASE_URL_ENV}=http://host:port" ) if not (url.startswith("http://") or url.startswith("https://")): raise BlogConfigError( f"{BASE_URL_ENV} must start with http:// or https:// (got {url!r})" ) ``` ```python # blog_manager/client.py:88-103 url = f"{self.base_url}{path}" resp = self.session.request( method, url, params=params, json=json, files=files, timeout=self.timeout, ) try: body: Any = resp.json() except ValueError: body = {"raw": resp.text} ``` ### Technical Analysis The `create-user` c ...[truncated 2050 chars]
Remediation
View remediation

T08 · Insecure Dependencies

Note
Location
requirements.txt:1
Finding

Open-Ended and Unhashed Dependency Specifications Allow Unreviewed Package Versions

Content
View full analysis
=2.28 pytest>=7.0 ``` The documented installation procedure executes these unconstrained requirements: ```bash pip install -r requirements.txt ``` ### Technical Analysis Both dependencies use open-ended minimum-version constraints. A future installation can therefore resolve to versions that did not exist when the project was audited and have not been reviewed or tested by the project maintainers. The requirements file also lacks cryptographic hashes, so the installer does not verify packages against an approved artifact set. In addition, `pytest`, which is a development and test dependency, is included in the same dependency file as the runtime HTTP library, unnecessarily increasing the production dependency surface. No evidence was found that either declared package is currently malicious. The weakness is that the installation process permits future or substituted artifacts outside a reviewed lock set. ### Attack Path 1. A user follows the documented command: ```bash pip install -r requirements.txt ``` 2. The package resolver selects the latest available versions satisfying the lower bounds. 3. If an upstream account, release, package index, mirror, or delivery path has been compromised, an unreviewed package artifact may be selected. 4. Installation or import of that package executes attacker-controlled code with the privileges of the user performing installation or running the CLI. This attack path is conditional on a supply-chain compromise or unsafe package source; the repository does not itself contain a confirmed malicious dependency. ### Impact Assessment A compromised dependency can execute with the privileges of the installing or invoking user. Depending on those privileges, it could read accessible files and environment variables, modify user ...[truncated 311 chars]
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (47)

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The description says the skill manages the Blog System API via 27 CLI subcommands. However, this code chunk does not actually implement those operations; it is a pytest test file that verifies expected CLI commands and related behavior. The tests strongly suggest the overall project does provide 27 commands, but the supplied chunk’s direct behavior is QA/validation, not API management. There is also a terminology inconsistency: the description says 'tags' while the tested command group is 'labels'. Finally, the code explicitly includes a meta command ('capability-list') and support behavior around formatting, environment configuration, and version hygiene that are not reflected in the declared description. Therefore the supplied code chunk is not accurately represented by the declared purpose.

Content

No source excerpt is available for this finding.

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · blog_manager/articles.py (reported line 8)May include surrounding context.

python
POST   /api/articles                   create_article
  GET    /api/articles/{id}              get_article
  PUT    /api/articles/{id}              update_article
  DELETE /api/articles/{id}?soft=        delete_article
  POST   /api/articles/{id}/restore      restore_article
  GET    /api/articles/heat/top          top_articles
"""

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · blog_manager/capability.py (reported line 26)May include surrounding context.

python
{"module": "articles", "name": "update-article",
     "description": "更新文章 (PUT /api/articles/{id})"},
    {"module": "articles", "name": "delete-article",
     "description": "删除文章,支持软删除 (DELETE /api/articles/{id})"},
    {"module": "articles", "name": "restore-article",
     "description": "恢复软删除的文章 (POST /api/articles/{id}/restore)"},
    {"module": "articles", "name": "top-articles",

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · blog_manager/capability.py (reported line 47)May include surrounding context.

python
{"module": "comments", "name": "list-comments",
     "description": "列出文章评论 (GET /api/comments/{aid})"},
    {"module": "comments", "name": "delete-comment",
     "description": "删除评论 (DELETE /api/comments/{comment_id})"},
    # Messages (4)
    {"module": "messages", "name": "list-messages",
     "description": "列出留言及回复 (GET /api/messages)"},

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · blog_manager/capability.py (reported line 56)May include surrounding context.

python
{"module": "messages", "name": "reply-message",
     "description": "回复留言 (POST /api/messages/reply)"},
    {"module": "messages", "name": "delete-message",
     "description": "删除留言 (DELETE /api/messages/{message_id})"},
    # Moods (3)
    {"module": "moods", "name": "list-moods",
     "description": "列出说说 (GET /api/moods)"},

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · blog_manager/capability.py (reported line 63)May include surrounding context.

python
{"module": "moods", "name": "create-mood",
     "description": "创建说说 (POST /api/moods)"},
    {"module": "moods", "name": "delete-mood",
     "description": "删除说说 (DELETE /api/moods/{mood_id})"},
    # Uploads (4)
    {"module": "uploads", "name": "upload-file",
     "description": "上传单个文件 (POST /api/upload, multipart file)"},

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · blog_manager/capability.py (reported line 72)May include surrounding context.

python
{"module": "uploads", "name": "list-uploads",
     "description": "列出已上传文件 (GET /api/uploads/list)"},
    {"module": "uploads", "name": "delete-upload",
     "description": "删除已上传文件 (DELETE /api/uploads/{filename})"},
    # Health (1)
    {"module": "health", "name": "health-check",
     "description": "健康检查 (GET /health)"},

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · blog_manager/comments.py (reported line 6)May include surrounding context.

python
Endpoints:
  POST   /api/comments            create_comment
  GET    /api/comments/{aid}      list_comments
  DELETE /api/comments/{id}       delete_comment
"""

from __future__ import annotations

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · blog_manager/messages.py (reported line 7)May include surrounding context.

python
GET    /api/messages               list_messages
  POST   /api/messages               create_message
  POST   /api/messages/reply         reply_message
  DELETE /api/messages/{id}          delete_message
"""

from __future__ import annotations

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · blog_manager/moods.py (reported line 6)May include surrounding context.

python
Endpoints:
  GET    /api/moods               list_moods
  POST   /api/moods               create_mood
  DELETE /api/moods/{id}          delete_mood
"""

from __future__ import annotations

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · blog_manager/uploads.py (reported line 7)May include surrounding context.

python
POST   /api/upload               upload_file    (multipart: file)
  POST   /api/upload/multiple      upload_files   (multipart: files)
  GET    /api/uploads/list         list_uploads
  DELETE /api/uploads/{filename}   delete_upload
"""

from __future__ import annotations

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
95% confidence
Finding

This endpoint deletes articles and is documented in an API that also states there is no authentication. Within a blog-manager skill, exposing a destructive operation without clear guardrails or authorization enables accidental or malicious deletion of content, including bulk abuse if automated by an agent.

Content

Scanner excerpt · references/api-reference.md (reported line 60)May include surrounding context.

md
响应:`{code: 200, message: "..."}`

### DELETE /api/articles/{article_id}

删除文章(支持软删除)。

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
94% confidence
Finding

This endpoint deletes comments with no documented authentication or ownership checks. In this skill context, an attacker or misused agent could censor discussion, remove evidence of abuse, or disrupt community interactions by deleting arbitrary comments.

Content

Scanner excerpt · references/api-reference.md (reported line 155)May include surrounding context.

md
响应:`{code: 200, data: [{id, uname, content, createtime}, ...]}`

### DELETE /api/comments/{comment_id}

删除评论。

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
94% confidence
Finding

The message deletion route is destructive and appears available in an unauthenticated API. That creates a straightforward path for unauthorized removal of user-submitted messages, undermining integrity and enabling harassment, censorship, or operational disruption.

Content

Scanner excerpt · references/api-reference.md (reported line 196)May include surrounding context.

md
响应:`{code: 200, data: {id}}`

### DELETE /api/messages/{message_id}

删除留言。

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
93% confidence
Finding

Deleting moods/posts is a state-changing operation in a documented unauthenticated API. In a content-management skill, this makes unauthorized content removal easy and increases the chance that an agent can be manipulated into destructive actions without sufficient validation.

Content

Scanner excerpt · references/api-reference.md (reported line 226)May include surrounding context.

md
响应:`{code: 200, data: {id}}`

### DELETE /api/moods/{mood_id}

删除说说。

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
97% confidence
Finding

Deleting uploaded files by filename is especially dangerous in an unauthenticated API because it can lead to loss of site assets, media, and potentially broader file-management abuse if filename handling is weak. In this blog-manager context, the endpoint directly affects stored content integrity and could be abused at scale by an agent or remote caller.

Content

Scanner excerpt · references/api-reference.md (reported line 264)May include surrounding context.

md
响应:`{code: 200, data: [{filename, url, type, size}, ...]}`

### DELETE /api/uploads/{filename}

删除已上传文件。

Env Variable Harvesting

High
Category
Data Exfiltration
Confidence
60% confidence
Finding

Code enumerates, copies, or searches environment variables for secrets. Bulk environment access can collect credentials unrelated to the skill's stated purpose.

Content

Scanner excerpt · tests/test_cli.py (reported line 272)May include surrounding context.

python
class TestCLISubprocess:
    def test_capability_list_via_subprocess(self):
        base_dir = os.path.join(os.path.dirname(__file__), "..")
        env = dict(os.environ, BLOG_MANAGER_BASE_URL="http://test:1234")
        result = subprocess.run(
            [sys.executable, "main.py", "capability-list"],
            capture_output=True, text=True, cwd=base_dir,

Env Variable Harvesting

High
Category
Data Exfiltration
Confidence
60% confidence
Finding

Code enumerates, copies, or searches environment variables for secrets. Bulk environment access can collect credentials unrelated to the skill's stated purpose.

Content

Scanner excerpt · tests/test_cli.py (reported line 283)May include surrounding context.

python
def test_missing_env_var_exits_2(self):
        base_dir = os.path.join(os.path.dirname(__file__), "..")
        env = dict(os.environ)
        env.pop("BLOG_MANAGER_BASE_URL", None)
        result = subprocess.run(
            [sys.executable, "main.py", "health-check"],

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The primary skill description is written in Chinese and the README does not indicate that users may choose another language or that the skill is intentionally restricted to a Chinese-speaking context. This can violate language/locale policy requirements when a skill implicitly enforces a language without opt-in.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding

The skill exposes shell, network, environment, and file capabilities but does not declare any explicit tool scope or permissions boundaries. Because the documented commands include file upload, delete operations, and arbitrary network targeting via BLOG_MANAGER_BASE_URL, the absence of constraints increases the risk of unintended destructive actions or misuse against internal or attacker-controlled endpoints.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The trigger phrases are broad enough that normal user conversation about blogs, comments, or blog systems could invoke the skill unintentionally. In this skill, accidental invocation matters because the available operations include creating, updating, deleting content, uploading files, and making unauthenticated network requests to a configured target.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The documentation advertises destructive and data-modifying actions such as delete-article, delete-comment, delete-message, delete-mood, and delete-upload without prominent safeguards or confirmation requirements. Since the target API is described as unauthenticated and the skill can perform irreversible changes, accidental or induced use could directly damage content or stored files.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The delete_article function performs a deletion request, including an irreversible hard-delete mode when soft=False, but the code provides no confirmation prompt, user-facing log/message, or inline warning comment/docstring about the destructive action. For code files, destructive operations should have some form of user disclosure unless clearly covered elsewhere in markdown, which is not present in this file.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

This code file contains many natural-language command descriptions in Chinese while the module docstring and identifiers are in English. That effectively imposes a locale on users consuming these descriptions, and there is no indication of user opt-in or documented region-specific justification.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This code file contains multiple hard-coded user-facing strings in Chinese, beginning with the empty-state message at L036 and continuing throughout the renderers. Because the skill forces a specific language in its natural-language output without user opt-in or a documented locale restriction, it violates the language/locale policy criteria.

Content

No source excerpt is available for this finding.