Security audit

KaleidoSphere

Security checks for vulnerabilities and agentic risk

Overview

The skill is narrowly framed for KaleidoSphere requests and does not contain hidden persistence, credential collection, mutation, or remote-code behavior, though its local validator is weaker than its written contract implies.

This skill is reasonable to install for a controlled KaleidoSphere integration, but do not treat its validator as a complete data-loss-prevention layer. Only use it with an already configured trusted transport, avoid placing sensitive data or raw BI rows in optional fields, and rely on host-side schema validation before any external dispatch.

A.I.G

Vulnerability Patterns

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
Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally

Findings (1)

T09 · Insecure Skill Coding Practices

Warning

Location: scripts/validate-request.mjs:36
Finding: Insufficient Validation of Sensitive and Structured Request Values
Content: View full analysis
500) fail('INVALID_OBJECTIVE'); else process.stdout.write(JSON.stringify({valid: true, action: request.action, authority: 'proposal-only'}) + '\n'); ``` ### Technical Analysis The validator attempts to reject unsafe content by serializing the entire `input` object and applying a keyword-based regular expression. It then checks only the allowed top-level property names for `discovery`, `plan`, and `preview` requests. This does not enforce types, formats, maximum lengths, or structural restrictions for the following accepted properties: - `discovery.field` - `discovery.value` - `plan.receiptId` - `preview.receiptId` Consequently, `value` and `receiptId` can contain neste ...[truncated 2500 chars]
Remediation: View remediation

SkillSpector

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
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep