Back to skill

Security audit

个人技能工坊

Security checks for vulnerabilities and agentic risk

Overview

This Chinese-language skill builder uses disclosed local file generation and validation scripts, and I found no hidden exfiltration, backdoor persistence, or deceptive behavior.

Install this only if you are comfortable with a Chinese-language skill that can create or update local skill files and run local Python checkers on directories you provide. Prefer a version-pinned release ZIP with checksum verification over the unpinned npx command, keep backups before using generated scripts, separate input and output directories, and avoid packaging credentials or private raw materials into shared skills.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Findings (71)

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The skill claims interactive creation/upgrading of personal skills from user material, but the described actual behavior focuses on offline structural audit, experiment evaluation, filesystem traversal, hash checks, and reference/symlink inspection. This is a serious trust mismatch because it may cause broad local-analysis behavior under the cover of a benign authoring assistant.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The skill claims interactive creation/upgrading of personal skills from user material, but the described actual behavior focuses on offline structural audit, experiment evaluation, filesystem traversal, hash checks, and reference/symlink inspection. This is a serious trust mismatch because it may cause broad local-analysis behavior under the cover of a benign authoring assistant.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The skill claims interactive creation/upgrading of personal skills from user material, but the described actual behavior focuses on offline structural audit, experiment evaluation, filesystem traversal, hash checks, and reference/symlink inspection. This is a serious trust mismatch because it may cause broad local-analysis behavior under the cover of a benign authoring assistant.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The skill claims interactive creation/upgrading of personal skills from user material, but the described actual behavior focuses on offline structural audit, experiment evaluation, filesystem traversal, hash checks, and reference/symlink inspection. This is a serious trust mismatch because it may cause broad local-analysis behavior under the cover of a benign authoring assistant.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The skill claims interactive creation/upgrading of personal skills from user material, but the described actual behavior focuses on offline structural audit, experiment evaluation, filesystem traversal, hash checks, and reference/symlink inspection. This is a serious trust mismatch because it may cause broad local-analysis behavior under the cover of a benign authoring assistant.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The skill claims interactive creation/upgrading of personal skills from user material, but the described actual behavior focuses on offline structural audit, experiment evaluation, filesystem traversal, hash checks, and reference/symlink inspection. This is a serious trust mismatch because it may cause broad local-analysis behavior under the cover of a benign authoring assistant.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The skill claims interactive creation/upgrading of personal skills from user material, but the described actual behavior focuses on offline structural audit, experiment evaluation, filesystem traversal, hash checks, and reference/symlink inspection. This is a serious trust mismatch because it may cause broad local-analysis behavior under the cover of a benign authoring assistant.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The skill claims interactive creation/upgrading of personal skills from user material, but the described actual behavior focuses on offline structural audit, experiment evaluation, filesystem traversal, hash checks, and reference/symlink inspection. This is a serious trust mismatch because it may cause broad local-analysis behavior under the cover of a benign authoring assistant.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The skill claims interactive creation/upgrading of personal skills from user material, but the described actual behavior focuses on offline structural audit, experiment evaluation, filesystem traversal, hash checks, and reference/symlink inspection. This is a serious trust mismatch because it may cause broad local-analysis behavior under the cover of a benign authoring assistant.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The skill claims interactive creation/upgrading of personal skills from user material, but the described actual behavior focuses on offline structural audit, experiment evaluation, filesystem traversal, hash checks, and reference/symlink inspection. This is a serious trust mismatch because it may cause broad local-analysis behavior under the cover of a benign authoring assistant.

Content

No source excerpt is available for this finding.

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 · scripts/run_anomaly_selftest.py (reported line 21)May include surrounding context.

python
ROOT = Path(__file__).resolve().parent.parent
FIX = ROOT / "assets" / "fixtures" / "anomaly"
SCRIPTS = ROOT / "scripts"
ENV = {**os.environ, "PYTHONUTF8": "1", "PYTHONIOENCODING": "utf-8"}  # children never inherit a console codepage

# (checker, case file, expected exit code, expected failing check/clause ids or None)
EXPECT = [

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

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

python
r"\s*[:=]\s*['\"]?[A-Za-z0-9_./+=-]{16,}"
    )),
)
TEXT_SUFFIXES = {".md", ".txt", ".json", ".yaml", ".yml", ".toml", ".env"}
IGNORED_HASH_DIRS = {"__pycache__", ".pytest_cache", ".mypy_cache"}
IGNORED_HASH_SUFFIXES = {".pyc", ".pyo"}

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · scripts/toolkit.py (reported line 322)May include surrounding context.

python
r"\s*[:=]\s*['\"]?[A-Za-z0-9_./+=-]{16,}"
    )),
)
TEXT_SUFFIXES = {".md", ".txt", ".json", ".yaml", ".yml", ".toml", ".env"}
IGNORED_HASH_DIRS = {"__pycache__", ".pytest_cache", ".mypy_cache"}
IGNORED_HASH_SUFFIXES = {".pyc", ".pyo"}

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

This markdown file is written entirely in Chinese, and there is no accompanying note that the skill is region-specific or that users may choose another language. Under the policy, forcing a specific language without opt-in is a natural-language policy violation unless the locale constraint is explicitly justified.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The entire skill description and usage instructions are presented only in Chinese, with no indication that users may choose another language or that the skill is intentionally limited to a Chinese-speaking audience. This can constitute a language/locale policy issue because it effectively forces a specific language without documented opt-in or justification.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
89% confidence
Finding

The README instructs users to run npx skills add ... without pinning the skills package to a specific version. This creates a supply-chain risk: a future malicious or compromised release of the package could be fetched and executed on the user's machine at install time. In the context of a skill installer, this is more dangerous because the command is presented as the primary installation path and may be copied verbatim by users.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding

The skill text explicitly instructs use of Python scripts, local file access, shell execution, and potentially network-capable tooling, but the skill metadata does not declare any tool scope or allowed-tools boundaries. This creates an authorization ambiguity where a host agent may execute broader capabilities than users expect, increasing the risk of over-privileged file access, command execution, or data exfiltration during skill use.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The natural-language instructions and description consistently require Chinese comprehension, but the file does not state that the skill is China/Chinese-specific or offer the user an opt-in or alternative language. This can violate language/locale policy because it effectively forces a specific language for all users.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The display name, short description, and default prompt are all written exclusively in Chinese, which indicates a language-specific user experience. There is no visible opt-in, alternate language option, or explanation that this skill is intentionally region-specific, so it may violate language/locale policy requirements.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

This JSON manifest contains all user-facing scenario descriptions and assertions in Chinese, but it does not indicate that the skill is intentionally limited to Chinese-speaking users or that language is configurable. For a manifest file, this can also make trigger/scope expectations less explicit to broader users and evaluators because the applicable language context is undocumented.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

Multiple case definitions require outputs such as '中文 Markdown', '简体中文', or '中文一句判定' across the suite, but the file does not provide an opt-in language selection mechanism or a documented justification for restricting the skill to Chinese. This is a natural-language locale policy concern because the manifest hardcodes a specific language in many user-facing behaviors.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

This markdown file uses a single forced language for all user-facing instructions and descriptions. Under the policy, language constraints should either provide user choice or be clearly justified as region-specific; neither is stated here.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

This manifest explicitly requires the skill output to be in "简体中文". Under the policy, forcing a specific language is a natural-language policy violation unless the skill offers a language choice or clearly documents a justified locale-specific constraint, which is not present here.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

This manifest uses Chinese identifiers and values such as "编号", "数量", "未知", and "备注" in configuration fields, which indicates a fixed language/locale assumption. Because SQP-3 applies to all file types and there is no visible opt-in, alternative locale support, or justification that the skill is region-specific, this appears to violate the language/locale policy requirement.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

This JSON manifest includes multiple natural-language requirement strings entirely in Chinese, such as the values on L10, L19, and L28. Because the file provides no indication that the skill is region-specific or that users can opt into this locale, it appears to force a specific language, which matches the language/locale policy violation criteria.

Content

No source excerpt is available for this finding.