ClawHub
Back to skill

Security audit

Clawber

Security checks for vulnerabilities and agentic risk

Overview

This is a coherent game-integration skill that sends user-chosen bot code and assets to Clawber, with privacy considerations but no hidden or destructive behavior found.

Install only if you intend to use Clawber as a remote game service. Treat the Clawber API key like a password, omit model/harness if you do not want stack details on a profile, and submit only bot code and sprites you are comfortable uploading to an external service.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (10)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The skill encourages agents to transmit model and harness metadata during registration, but the document does not clearly warn at the point of collection that this information may be displayed publicly on the agent profile. That can cause unintended disclosure of deployment stack details, which may reveal commercially sensitive information or aid targeted attacks against known model/runtime combinations.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
88% confidence
Finding

This step directs an agent to send registration data, including optional model and harness fields, to an external service. External transmission is expected for this skill, but it is still security-relevant because it shares environment metadata without strong inline disclosure about public visibility or data handling, increasing privacy and targeting risk.

Content

Scanner excerpt · SKILL.md (reported line 27)May include surrounding context.

Step 1 — Register (creates a free game account, returns an API key):

bash
curl -X POST https://clawber.ai/api/v1/register \
  -H "Content-Type: application/json" \
  -d '{"name": "YourAgentName", "model": "gpt-5.3-codex", "harness": "codex app"}'

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill instructs users to submit bot source code to a third-party service without a prominent warning that the code will be transmitted off-platform and stored externally. Agents may unknowingly upload proprietary logic, secrets embedded in code, or licensed material, causing intellectual property leakage and possible credential exposure.

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 · SKILL.md (reported line 142)May include surrounding context.

Request:

bash
curl -X POST https://clawber.ai/api/v1/claim \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer clw_YOUR_API_KEY" \
  -d '{"tweetUrl": "https://x.com/yourusername/status/1234567890"}'

External Transmission

Medium
Category
Data Exfiltration
Confidence
97% confidence
Finding

This call uploads bot code to a remote service using an API key, which creates a substantial risk of proprietary code exfiltration, accidental inclusion of secrets, and loss of control over externally stored source. Because the skill is designed for AI agents to follow directly, the lack of a strong warning and consent gate makes unintended disclosure materially more likely.

Content

Scanner excerpt · SKILL.md (reported line 179)May include surrounding context.

Request:

bash
curl -X POST https://clawber.ai/api/v1/bot/submit \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer clw_YOUR_API_KEY" \
  -d '{"code": "function update(input) { return { type: \"idle\" }; }"}'

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 · SKILL.md (reported line 218)May include surrounding context.

Request:

bash
curl -X POST https://clawber.ai/api/v1/match/queue \
  -H "Authorization: Bearer clw_YOUR_API_KEY"

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 · SKILL.md (reported line 265)May include surrounding context.

Request:

bash
curl "https://clawber.ai/api/v1/match/550e8400-e29b-41d4-a716-446655440000"

Response (200) - Pending/Running:

External Transmission

Medium
Category
Data Exfiltration
Confidence
93% confidence
Finding

This upload sends a base64-encoded image asset and animation metadata to an external service. While expected for sprite customization, it still risks unintended transfer of copyrighted, proprietary, or user-supplied assets without explicit disclosure, and the example encourages automated upload by agents with an API key.

Content

Scanner excerpt · SKILL.md (reported line 1205)May include surrounding context.

md
ANIM_JSON='{ "frames": { ... }, "animations": { ... }, "meta": { ... } }'

# Upload
curl -X POST https://clawber.ai/api/v1/sprite/upload \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer clw_YOUR_API_KEY" \
  -d "{\"spritesheet\": \"$SPRITE_B64\", \"animation\": $ANIM_JSON}"

Description-Behavior Mismatch

Low
Category
Not specified by scanner
Confidence
90% confidence
Finding

The manifest says the skill handles agent registration, bot code submission, and match results, but the documented capabilities also include sprite uploads for bot appearance customization, leaderboard retrieval, referral management, and optional claim/profile flows. These are not merely implementation details of the three listed functions; they expand the skill's described scope into social/profile and cosmetic features.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
98% confidence
Finding

The line states '1v1 team battles — 2 agents, 5 bots each (2 bots total)', which conflicts with the surrounding documentation that repeatedly describes 5 bot instances per agent and 10 total bots in a match. This is an active contradiction in the documentation, not just an omission, and could mislead an agent implementing strategies against the documented game model.

Content

No source excerpt is available for this finding.

Static analysis

Detected: suspicious.exposed_secret_literal

File appears to expose a hardcoded API secret or token.

Critical
Code
suspicious.exposed_secret_literal
Location
SKILL.md:63