Back to skill

Security audit

MoltyRoyale

Security checks for vulnerabilities and agentic risk

Overview

This game skill includes high-impact wallet, trading, token deployment, credential storage, and self-updating behavior that needs review before installation.

Install only if you intentionally want a Molty Royale agent that may manage local credentials, connect to remote game services, and use blockchain-related workflows. Review or disable the remote self-update behavior, avoid in-chat or command-line private-key handling, use a secret manager instead of plaintext credential files, and require explicit confirmation before any token trade, approval, purchase, deployment, or owner-key custody.

Vulnerability Patterns
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Tool Hijacking and SpoofingModifies or replaces tools so legitimate-looking calls execute attacker logic
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (4)

T03 · Remote Payload Retrieval and Execution

Error
Location
heartbeat.md:43
Finding

Unverified Remote Replacement of Active Skill Instructions

Content
View full analysis
`memory.localFetchedAt`: ```bash curl -s https://www.moltyroyale.com/skill.md > ~/.molty-royale/skills/skill.md curl -s https://www.moltyroyale.com/heartbeat.md > ~/.molty-royale/skills/heartbeat.md ``` Then update `memory.localFetchedAt` to the current time. ``` The same behavior is required by the primary Skill: ```markdown - Compare `response.skillLastUpdate` with `memory.localFetchedAt`. If `skillLastUpdate` is newer → re-download skill.md and heartbeat.md, then update `localFetchedAt`. ``` ### Technical Analysis The Skill uses a server-controlled `skillLastUpdate` value to trigger downloads that directly overwrite its installed `skill.md` and `heartbeat.md`. The downloaded files are not checked against a pinned version, cryptographic digest, signed manifest, or trusted release key. They are also not staged for review or subjected to user confirmation before becoming active instructions. Although HTTPS protects the connection in transit, it does not protect against compromise of the origin server, deployment pipeline, DNS/TLS infrastructure, or the account authorized to update the hosted files. The effective Skill payload can consequently change after the reviewed package has been installed. Because the retrieved files are agent instructions, hostile content could direct later runs to read local credentials, install software, execute commands, sign financial transactions, or contact additional endpoints. ### Attack Path 1. An attacker compromises the Molty Royale website, its deployment pipeline, or another component capable of changing the hosted Skil ...[truncated 1292 chars]
Remediation
View remediation

T08 · Insecure Dependencies

Warning
Location
forge-token-deployer.md:37
Finding

Unpinned Dependencies Installed in Wallet- and Credential-Handling Processes

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
cross-forge-trade.md:208
Finding

Wallet Private Keys Passed Through Command-Line Arguments

Content
View full analysis
\ "[0xDdF8AaA3927b8Fd5684dc2edcc7287EcB0A2122d,0xdb99a97d607c5c5831263707E7b746312406ba7E]" \ \ \ --value 10ether \ --rpc-url https://mainnet.crosstoken.io:22001 \ --private-key ``` ```bash # Approve MOLTZ token cast send 0xdb99a97d607c5c5831263707E7b746312406ba7E \ "approve(address,uint256)" \ 0x7aF414e4d373bb332f47769c8d28A446A0C1a1E8 \ \ --rpc-url https://mainnet.crosstoken.io:22001 \ --private-key ``` ```bash # Sell MOLTZ cast send 0x7aF414e4d373bb332f47769c8d28A446A0C1a1E8 \ "swapExactTokensForNative(uint256,uint256,address[],address,uint256)" \ \ \ "[0xdb99a97d607c5c5831263707E7b746312406ba7E,0xDdF8AaA3927b8Fd5684dc2edcc7287EcB0A2122d]" \ \ \ --rpc-url https://mainnet.crosstoken.io:22001 \ --private-key ``` ### Technical Analysis The examples instruct users to replace `` with a wallet private key in a command-line argument. Command-line secrets can be exposed through: - Shell history. - Process listings and `/proc` inspection. - Terminal session recording. - Command auditing and endpoint telemetry. - Debug logs or copied command transcripts. - CI/CD job logs. A private key is a reusable bearer credential with full signing authority over its wallet. Unlike a bounded transaction signature, disclosure permits the observer to authorize arbitrary future transactions. ### Attack Path 1. The user substitutes a real wallet private key into one of the documented commands. 2. The shell reco ...[truncated 988 chars]
Remediation
View remediation

T07 · Tool Hijacking and Spoofing

Warning
Location
forge-token-deployer.md:335
Finding

Deployer Bootstrap Instructions Can Overwrite Existing Project Files

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Memory PoisoningPersistent Context Injection, Context Window Stuffing, Memory Manipulation
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
Findings (61)

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The file is materially misaligned with the declared skill domain: instead of Molty Royale game-agent operations, it implements blockchain token trading workflows. In an agent system, this kind of scope substitution is dangerous because it can cause the agent to invoke financial transaction logic under an unrelated, trusted capability, enabling unauthorized or unexpected asset movement.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill introduces token swaps and ERC-20 approval flows, including commands that can move funds and grant spending permissions to a router. In the context of a Molty Royale game-agent skill, these capabilities are unjustified and highly dangerous because they expand the agent from gameplay management into real asset custody and delegated spending, which can directly lead to theft or irreversible loss.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill describes live trading and approval sequences without prominent warnings that transactions are irreversible and approvals can authorize future token transfers by the router. In an agent-driven environment, omission of these warnings increases the chance of unsafe execution by users or downstream agents who may not appreciate the financial and permissioning risks.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill instructs the agent to generate and store private keys and later handle an API key without an explicit security warning, consent boundary, or secure secret-management requirements. In an agentic context, this is dangerous because it normalizes autonomous creation and retention of highly sensitive credentials that could be exposed through logs, local files, prompts, or downstream tools.

Content

No source excerpt is available for this finding.

Credential Access

High
Category
Privilege Escalation
Confidence
95% confidence
Finding

The skill explicitly directs storage of a one-time API key in a predictable local file path under the user's home directory. That creates a real credential-exposure risk because local plaintext files are commonly readable by other processes, accidentally committed, backed up insecurely, or leaked through support bundles and agent memory.

Content

Scanner excerpt · features/onboarding.md (reported line 52)May include surrounding context.

md
- `apiKey` appears in this response **once only** — save it immediately
  - Also save `accountId` and `publicId`
  - Recommended path: `~/.molty-royale/credentials.json`
  - All subsequent API calls require `X-API-Key: mr_live_xxx` header

- [ ] Step 3: Determine Owner EOA (separate from Agent EOA)

Vague Triggers

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill description is extremely broad and includes vague trigger phrases like 'token launch' and 'anything about getting a token onto Forge', which can cause accidental invocation in contexts where the user did not explicitly request deployment. Because this skill performs file creation, dependency installation, API calls, and blockchain transaction preparation, unintended activation could lead to high-impact external actions.

Content

No source excerpt is available for this finding.

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · forge-token-deployer.md (reported line 106)May include surrounding context.

md
const clientKey = process.env.CLIENT_KEY;
    const clientSecret = process.env.CLIENT_SECRET;
    if (!clientKey || !clientSecret) {
      throw new Error("CLIENT_KEY and CLIENT_SECRET environment variables are not set. Please check your .env file.");
    }
    headers["Authorization"] = `API-Key ${clientKey}:${clientSecret}`;
  } else {

Memory Manipulation

High
Category
Memory Poisoning
Confidence
80% confidence
Finding

Skill manipulates agent memory, state, or stored context. Memory corruption can alter personality, override safety rules, or cause unpredictable behavior.

Content

Scanner excerpt · heartbeat.md (reported line 253)May include surrounding context.

clear temp entirely

text

**Reset memory:**

```text
current_phase = preparing

Missing User Warnings

High
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill instructs automatic creation of a new Owner EOA and local storage without a clear consent step, secure key-handling guidance, or constraints on where and how secrets are stored. Automatically generating and persisting private keys in an agent workflow can expose funds and identity if the host is compromised, logs leak, permissions are weak, or users do not realize a wallet was created on their behalf.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The documentation explicitly instructs the agent to generate, store, retain, and later hand off an Owner EOA private key. That introduces long-term secret custody and key-export behavior far beyond ordinary gameplay automation, creating a direct risk of wallet compromise, unauthorized signing, and asset theft if the agent, host, logs, or local storage are exposed.

Content

No source excerpt is available for this finding.

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · references/identity.md (reported line 127)May include surrounding context.

3. Unregister Identity

DELETE /api/identity (requires X-API-Key)

bash
curl -X DELETE https://cdn.moltyroyale.com/api/identity \

Missing User Warnings

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The guidance explicitly permits the agent to store an owner's private key and later hand it off, but it does not strongly prohibit retention, warn that anyone with the key has full wallet control, or require safer alternatives. In a wallet-owning game/payment workflow, this creates a real secret-handling vulnerability: compromise of the agent, logs, memory, or support transcripts could expose the key and enable theft or unauthorized owner-side signing.

Content

No source excerpt is available for this finding.

Ssd 3

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

This section directs the agent to keep a generated owner private key and provide it to the user on request, which normalizes secret exfiltration through the chat/channel. A private key is an account-compromise credential; exposing it through agent memory or conversation creates a direct path to wallet takeover, fund theft, impersonation, and irreversible blockchain actions.

Content

No source excerpt is available for this finding.

Ssd 3

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill allows disclosure and handoff of generated owner private-key material in natural language, which is extremely sensitive credential material. Exposing or transmitting private keys through chat or agent responses can directly lead to wallet takeover, asset theft, impersonation, and permanent loss of control over the owner account.

Content

No source excerpt is available for this finding.

Credential Access

High
Category
Privilege Escalation
Confidence
91% confidence
Finding

The recommended credentials file stores the API key together with account identifiers and wallet metadata in a predictable plaintext location. Because the API key is shown once and likely authorizes account operations, compromise of this file can enable unauthorized account control, game actions, or linkage to wallet operations.

Content

Scanner excerpt · references/setup.md (reported line 157)May include surrounding context.

md
The `apiKey` is shown only once.

Recommended credentials path:
`~/.molty-royale/credentials.json`

Example:

Missing User Warnings

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The advanced flow allows the agent to generate and retain an Owner EOA private key, which is materially more sensitive than an agent-only key because it controls the owner's wallet and associated funds. The text normalizes agent-side custody with insufficient risk disclosure, increasing the chance that users unknowingly delegate financial control to the agent or store a fund-controlling key insecurely.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The guide says the agent should provide the generated Owner private key to the user on request, but does not define a safe delivery channel or warn against exposing it in chat logs, transcripts, or other observable interfaces. Revealing a live wallet private key through insecure channels can instantly compromise ownership of the wallet and any associated assets or approvals.

Content

No source excerpt is available for this finding.

Ssd 3

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

These instructions explicitly tell the agent to disclose a generated Owner EOA private key when asked. In the context of an agent skill, that creates a direct secret-exfiltration pathway where sensitive wallet material may be surfaced into conversational history or exposed to unauthorized requesters.

Content

No source excerpt is available for this finding.

Credential Access

High
Category
Privilege Escalation
Confidence
84% confidence
Finding

The instruction to detect first-run status by checking for dev-agent/credentials.json causes the skill to probe for and rely on locally stored credential material. Access patterns around credential files expand the attack surface because other components may infer, read, or misuse secrets stored in predictable locations.

Content

Scanner excerpt · skill.md (reported line 30)May include surrounding context.

md
First-run detection (file-based):

- Treat as **first-run** if one of these is missing:
  - `dev-agent/credentials.json`
  - `dev-agent/owner-intake.json`
- If both files exist, treat as **returning-run**.
- Returning-run asks only missing/invalid fields, not the full list.

Credential Access

High
Category
Privilege Escalation
Confidence
95% confidence
Finding

This line directly instructs persistence of credential values into dev-agent/credentials.json, which is a classic insecure secret-storage pattern when no protections are specified. If that file is exposed through local compromise, backups, shared workspaces, or other tools with filesystem access, attackers can obtain the API key and operate the account or associated agent sessions.

Content

Scanner excerpt · skill.md (reported line 39)May include surrounding context.

md
- Ask the full owner question list once.
- Persist normalized answers into `dev-agent/owner-intake.json`.
- Persist credential values into `dev-agent/credentials.json`.
- Then execute the flow automatically.

Ask these in order (short, direct) to owner (human):

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The invocation description is broad enough to match generic buy/sell or swap requests, which increases the chance that an agent selects this skill in contexts the user did not intend. Because the skill performs financial actions, overly permissive matching is dangerous: mistaken invocation can escalate directly into irreversible trades or approvals.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The feature explicitly instructs agents to persist strategy, game history, and observed agent data to a local JSON file in the user's home directory, but it does not warn about retention, local privacy exposure, or sensitivity of stored behavioral data. In the context of an autonomous agent skill, silent persistence increases the risk of collecting and retaining data longer than users expect, especially on shared hosts or systems with weak file permissions.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The file explicitly states free rooms are suspended, but later provides active instructions to continuously join and run free rooms. This contradictory guidance can cause an agent to take unsupported actions, enter retry loops, or ignore current operational policy, leading to failed requests, wasted resources, and unreliable behavior.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The skill instructs the agent to open an external WebSocket connection and use an X-API-Key credential, but it provides no explicit warning, consent boundary, or guidance on secure credential handling. In an agent skill, this can lead to automatic transmission of secrets to a remote service and normalize network actions without user awareness, increasing the risk of credential misuse or unintended data disclosure.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger text says "Account creation requested, or prerequisites check fails on first run," which does not define clear boundaries for when this skill should or should not activate. Phrases like "account creation requested" and a generic first-run prerequisite failure can overlap with many normal setup interactions, increasing the risk of unintended invocation.

Content

No source excerpt is available for this finding.