TinkerClaw Model Router

• Outcome: Pass
• Latest audit: Sep 17, 2026 · 1:18 PM UTC
• Version: 2.0.2

Overview

This skill is an instruction-only guide for choosing cheaper or stronger configured models, with no hidden code or automatic data sharing found.

Review the routing preferences before installing because they may influence which configured model handles future agent or cron work. The persistent AGENTS.md snippet should only be added if you want this routing policy to apply across sessions.

A.I.G

Vulnerability Patterns

• 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
• Unauthorized Access and Privilege Escalation: Obtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• 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 (3)

Whitespace Padding

Medium

• Category: Prompt Injection
• Confidence: 70% confidence

Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · SKILL.md (reported line 173)May include surrounding context.

md

### Seven-Day Window (flat-rate provider quota)

| seven_day utilization | Routing guidance                                                                                                             |
| --------------------- | ---------------------------------------------------------------------------------------------------------------------------- |
| < 70%                 | **Normal** — best model for the task tier                                                                                    |
| 70–85%                | **Conservative** — prefer sonnet over opus; haiku for flash/fast tiers                                                       |

Whitespace Padding

Medium

• Category: Prompt Injection
• Confidence: 70% confidence

Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · SKILL.md (reported line 175)May include surrounding context.

md

| seven_day utilization | Routing guidance                                                                                                             |
| --------------------- | ---------------------------------------------------------------------------------------------------------------------------- |
| < 70%                 | **Normal** — best model for the task tier                                                                                    |
| 70–85%                | **Conservative** — prefer sonnet over opus; haiku for flash/fast tiers                                                       |
| 85–95%                | **Austerity** — sonnet max for everything except active user conversation with explicit escalation. No opus for crons.       |
| > 95%                 | **Emergency** — haiku + local only. Opus only for active user conversation with explicit ask. No automated work above haiku. |

Natural-Language Policy Violations

Low

• Category: Not specified by scanner
• Confidence: 85% confidence

Finding

The decision tree routes all translation tasks to a single model choice without any mention of preserving the user's language or offering locale selection. In a reference document that may guide downstream skill behavior, this is a natural-language policy concern because it can encourage language handling without explicit user opt-in or locale consideration.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.