AI Product Comparison Skill

• Outcome: Review
• Latest audit: Sep 16, 2026 · 4:37 AM UTC
• Version: 1.0.0

Overview

The skill is mostly a coherent product-comparison integration, but it includes unsafe troubleshooting and credential/file-handling practices that warrant review before installation.

Review before installing. Use a dedicated Zyte API key with limited quota, do not store real keys in shared or committed config files, avoid authenticated or sensitive URLs, and do not run the documented /etc/resolv.conf command. Raw extraction files may be left in predictable /tmp paths, so use this only in a trusted local environment or revise the scripts first.

A.I.G

Vulnerability Patterns

• Unauthorized Access and Privilege Escalation: Obtains permissions beyond the task's legitimate needs
• Insecure Skill Coding Practices: Finds exploitable flaws such as hardcoded secrets or command injection
• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review

Findings (5)

T09 · Insecure Skill Coding Practices

Warning

• Location: scripts/fetch_products.py:238

Finding

API Key Exposed Through Process Command-Line Arguments

Content

View full analysis

Remediation

View remediation

T09 · Insecure Skill Coding Practices

Note

• Location: SKILL.md:73

Finding

API Key Prefix Disclosed During Environment Validation

Content

View full analysis

Remediation

View remediation

T09 · Insecure Skill Coding Practices

Error

• Location: scripts/fetch_products.py:99

Finding

Predictable Shared Temporary Files Permit Symlink Attacks and Data Exposure

Content

View full analysis

Remediation

View remediation

T05 · Unauthorized Access and Privilege Escalation

Warning

• Location: SKILL.md:278

Finding

Troubleshooting Instructions Overwrite Global DNS Configuration

Content

View full analysis

/etc/resolv.conf ``` ### Technical Analysis The instruction replaces the host's entire resolver configuration with a single public DNS server. Product comparison does not legitimately require a persistent, system-wide DNS modification. Writing `/etc/resolv.conf` normally requires elevated privileges. In a privileged agent environment, the command can discard corporate resolvers, search domains, split-DNS settings, local stub resolvers, VPN configuration, and security filtering. It also routes DNS queries to an external provider. The command performs no backup, confirmation, scope restriction, compatibility check, or restoration. ### Attack Path 1. A network or sandbox DNS failure occurs. 2. The agent follows the Skill's troubleshooting instruction in an environment with sufficient privileges. 3. Existing `/etc/resolv.conf` content is overwritten. 4. The host begins using the specified third-party resolver. 5. Unrelated applications experience policy bypass, resolution failures, or DNS-query disclosure. ### Impact Assessment The instruction may affect all processes using the host resolver, not only this Skill. Potential consequences include: - Loss of organization-specific name resolution - Disruption of VPN, container, or local resolver behavior - DNS-query disclosure to a third party - Bypass of enterprise DNS filtering or logging - Network outages for unrelated applications It does not grant new privileges by itself, but it directs an already privileged agent to make a change beyond the minimum privileges required for the declared task. ]]>

Remediation

View remediation

T09 · Insecure Skill Coding Practices

Note

• Location: README.md:117

Finding

Documentation Encourages Plaintext API-Key Persistence

Content

View full analysis

Remediation

View remediation

SkillSpector

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration

Findings (11)

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 96% confidence

Finding

The declared description presents a higher-level product comparison skill: extracting structured product data, building side-by-side comparison tables, and providing intelligent purchase recommendations. The supplied code chunk only implements the fetch stage. It sends Zyte API requests with {"url": ..., "product": true}, handles retries/rate limits, saves raw API responses to /tmp/product__raw.json, and prints a machine-readable fetch summary. While this partially aligns with the extraction portion of the description, it does not itself perform structured parsing, cross-product comparison, table generation, or recommendation logic. Therefore the code's actual behavior is materially narrower than the declared primary purpose.

Content

No source excerpt is available for this finding.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 97% confidence

Finding

The description presents a higher-level product comparison skill: ingest multiple e-commerce URLs, use the Zyte API, compare products, and recommend which to buy. The actual code chunk is much narrower. It only parses an already-saved Zyte API response from a local file into normalized JSON for one product. This parsing step could be a supporting component of the broader skill, but by itself it does not implement the core declared behaviors—API extraction from URLs, multi-product comparison, comparison tables, or purchase recommendations. Therefore the supplied code chunk does not accurately represent the declared purpose.

Content

No source excerpt is available for this finding.

Missing User Warnings

High

• Category: Not specified by scanner
• Confidence: 99% confidence

Finding

The markdown recommends a destructive privileged system change without warning about its scope or side effects. Overwriting /etc/resolv.conf can break local name resolution, alter security monitoring and policy enforcement, and affect unrelated applications and sessions beyond the skill's execution.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium

• Category: Not specified by scanner
• Confidence: 96% confidence

Finding

The README describes extracting product data from user-provided URLs via the Zyte API but does not prominently warn that those URLs will be transmitted to a third-party service. Even if the URLs are public pages, they may contain sensitive query strings, tracking tokens, private storefront paths, or business-sensitive browsing intent, so silent sharing creates a privacy and data-handling risk.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium

• Category: Not specified by scanner
• Confidence: 95% confidence

Finding

The README says the skill triggers automatically when it detects product-comparison intent and e-commerce URLs, but does not define clear activation boundaries or exclusions. In an agent environment, ambiguous triggers can cause unintended invocation on loosely related prompts or pasted URLs, leading to unnecessary outbound requests, surprise actions, and possible disclosure of user-supplied links to the external service.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium

• Category: Not specified by scanner
• Confidence: 98% confidence

Finding

The README tells users to overwrite /etc/resolv.conf with nameserver 8.8.8.8, which is a privileged system-wide network configuration change unrelated to the core product-comparison function. This can disrupt enterprise DNS, break VPN/split-DNS setups, weaken monitoring or policy enforcement, and trains users to run unsafe commands copied from troubleshooting docs.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium

• Category: MCP Least Privilege
• Confidence: 95% confidence

Finding

The skill instructs the agent to perform network access and write response files under /tmp, but it does not declare any explicit tool scope or permissions boundary. That increases the risk of over-broad execution because a host system may allow capabilities beyond what the skill actually needs, making misuse or accidental abuse harder to constrain and audit.

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 overly broad and can cause the skill to activate for common shopping-advice requests that may not require external fetching or API use. Overbroad activation increases the chance of unnecessary network access, secret use via ZYTE_API_KEY, and unintended processing of user-provided URLs, especially when the skill also has side effects like writing files and making remote calls.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium

• Category: Not specified by scanner
• Confidence: 99% confidence

Finding

The skill tells the operator to overwrite /etc/resolv.conf with a public DNS server. That is a privileged system configuration change unrelated to the core business logic of product comparison, and it can disrupt host networking, bypass enterprise DNS controls, and create confidentiality and integrity risks for all subsequent network traffic on the system.

Content

No source excerpt is available for this finding.

External Transmission

Medium

• Category: Data Exfiltration
• Confidence: 60% confidence

Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · scripts/fetch_products.py (reported line 118)May include surrounding context.

python

}

    req = urllib.request.Request(
        "https://api.zyte.com/v1/extract",
        data=payload,
        headers=headers,
        method="POST",

Scope Creep

Low

• Category: Excessive Agency
• Confidence: 70% confidence

Finding

Skill's behavior or capabilities extend beyond its stated purpose. Scope creep allows an agent to perform actions unrelated to its documented functionality, increasing the attack surface.

Content

Scanner excerpt · LICENSE.md (reported line 5)May include surrounding context.

md

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

Static analysis

No suspicious patterns detected.