Security audit

Climate

Security checks for vulnerabilities and agentic risk

Overview

The skill presents itself as a climate toolkit, but the included script mainly stores arbitrary user inputs in local logs and can replay or export them.

Install only if you intentionally want a local command-input logging utility under ~/.local/share/climate. Do not enter secrets, credentials, private notes, or sensitive business data; review and clear the directory if you test it. The artifact does not show exfiltration or destructive behavior, but its purpose and retention behavior are under-scoped for a climate skill.

A.I.G

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

SkillSpector

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration

Findings (8)

Tp4

High

Category: MCP Tool Poisoning
Confidence: 97% confidence
Finding: The skill is presented as a simple climate-related command-line tool, but the documentation also describes persistent local storage, activity logging, and export/search capabilities that are broader than the stated purpose. This creates a trust and transparency problem: users may invoke the skill expecting a lightweight informational tool while it writes data and retains logs on disk without clearly declaring those side effects up front.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The manifest description says "Use when you need climate," which does not define a clear trigger scope or activation boundary. This is vague enough to overlap with ordinary requests and gives no constraints or negative examples to distinguish when the skill should or should not be invoked.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 91% confidence
Finding: The documentation states that the tool stores data locally and logs command activity with timestamps, but it does not warn users about these data side effects, retention, or possible sensitivity of logged activity. This is dangerous because users may unknowingly create persistent records of their actions or data, which could later be exposed through local compromise, backups, shared systems, or exports.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 91% confidence
Finding: The inline documentation frames the script as a benign utility while the code's core behavior is persistent logging of arbitrary command inputs. Deceptive or inaccurate documentation is dangerous because it reduces user scrutiny and can lead to disclosure of secrets or sensitive operational data to local storage.
Content: No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium

Category: Not specified by scanner
Confidence: 96% confidence
Finding: The script creates a persistent data directory under the user's home directory and appends user-provided content to log files without any minimization or consent. If users enter credentials, API keys, internal notes, or personal data, that information will remain on disk and may later be accessed by other local processes, backups, exports, or anyone with access to the account.
Content: No source excerpt is available for this finding.

Ssd 3

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The tool not only records arbitrary input but also re-exposes it through status, recent activity, search, and export features, expanding the chance of secondary disclosure. In this skill context, the stated 'climate' purpose does not justify broad collection and replay of user-provided content, making the behavior more suspicious and more dangerous because users are less likely to expect sensitive data handling.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The advertised functionality is misleading: the help text presents a generic 'Climate' utility, but the implementation primarily records arbitrary user inputs into persistent local logs. This mismatch can cause users or higher-level agents to provide sensitive data under false assumptions, increasing the risk of unintended collection and retention.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: These command handlers write raw user input directly to per-command log files and history without warning the user that their data is being stored. Silent persistence is risky because users may treat the tool like a transient CLI and unknowingly disclose sensitive information that is retained indefinitely.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.