Weather Like Test

• Outcome: Pass
• Latest audit: Sep 10, 2026 · 6:18 PM UTC
• Version: 0.1.0

Overview

This is a simple weather lookup skill with expected external weather-service calls, but users should avoid precise private locations and prefer HTTPS examples.

Use coarse locations such as city or airport code rather than exact home/work coordinates. Prefer adding explicit https:// to wttr.in commands before use, especially on untrusted networks.

A.I.G

Vulnerability Patterns

• 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
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally

Findings (1)

T09 · Insecure Skill Coding Practices

Warning

• Location: SKILL.md:15

Finding

Plaintext HTTP Requests Expose Weather Queries to Network Interception

Content

View full analysis

Vulnerability Details

File Location: SKILL.md, lines 15–37
Vulnerability Type: Plaintext HTTP communication
Risk Level: Medium

The following documented commands access wttr.in without explicitly requiring HTTPS:

bash

curl -s "wttr.in/London?format=3"

bash

curl -s "wttr.in/London?format=%l:+%c+%t+%h+%w"

bash

curl -s "wttr.in/London?T"

bash

curl -s "wttr.in/Berlin.png" -o /tmp/weather.png

Technical Analysis

URLs without an explicit scheme are interpreted by curl as HTTP URLs. Consequently, the initial weather request is transmitted over plaintext HTTP. A server-side redirect to HTTPS does not protect the initial request because the requested location and HTTP headers have already crossed the network without transport encryption.

An attacker able to intercept network traffic could observe requested locations or modify the plaintext response. The attacker could also return a malicious redirect or substitute the downloaded weather image. No evidence indicates that the downloaded image is executed or processed by a privileged component, so arbitrary code execution is not established.

Attack Path

1. A user or Agent runs one of the documented wttr.in commands.
2. Curl initiates an unencrypted HTTP connection because the URL does not specify https://.
3. An attacker positioned on the same untrusted network, or another party capable of intercepting the route, observes or modifies the request.
4. The attacker learns the queried location, alters weather output, injects a redirect, or substitutes the downloaded PNG response.
5. The user or Agent receives attacker-controlled or misleading content.

Impact Assessment

Exploitation requires a network-adjacent or otherwise on-path attacker. It may disclose location queries and compromise the confidentiality and integrity of weather responses. The affected scope is limited to traffic generated by these wttr.i ...[truncated 129 chars]

Remediation

View remediation

Remediation Suggestions

Replace every wttr.in reference with an explicit HTTPS URL. Configure curl to reject non-HTTPS protocols, report HTTP failures, and follow only secure redirects where redirects are necessary.

Recommended examples:

bash

curl --fail --show-error --silent --location \
  --proto '=https' --proto-redir '=https' \
  "https://wttr.in/London?format=3"

bash

curl --fail --show-error --silent --location \
  --proto '=https' --proto-redir '=https' \
  "https://wttr.in/Berlin.png" \
  --output /tmp/weather.png

Apply the same explicit https:// scheme and protocol restrictions to all other wttr.in examples. If downloaded files are later consumed by another component, validate their media type and content before use.

SkillSpector

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (2)

External Transmission

Medium

• Category: Data Exfiltration
• Confidence: 86% confidence

Finding

The documentation directs users to transmit latitude and longitude to api.open-meteo.com, which is a third-party external service. In context this is functionally necessary for the skill, but coordinates can be sensitive and may reveal a user's exact location if copied from a private place.

Content

Scanner excerpt · SKILL.md (reported line 44)May include surrounding context.

Free, no key, good for programmatic use:

bash

curl -s "https://api.open-meteo.com/v1/forecast?latitude=51.5&longitude=-0.12&current_weather=true"

Find coordinates for a city, then query. Returns JSON with temp, windspeed, weathercode.

Missing User Warnings

Low

• Category: Not specified by scanner
• Confidence: 89% confidence

Finding

The skill instructs users to send location queries to third-party weather services without clearly warning that the requested city, airport code, or coordinates will be transmitted externally. While this is expected for a weather skill and not inherently malicious, it can expose sensitive location information if users query home addresses, precise coordinates, or other private places.

Content

No source excerpt is available for this finding.