Most static alerts are false positives caused by Markdown fences, JavaScript template literals, expected API calls, and documented credential configuration. Two confirmed findings show credentials in terminal output, and the agent workflow lacks a boundary for untrusted public ideaboard content. The author should prevent secret printing and require review of public content before acting on it.
When no wallet key is configured, the example prints a newly generated private key in plaintext. This can expose control of the generated wallet through logs or terminal history.
Untrusted Ideaboard Content Can Influence Agent Decisions
The workflow instructs agents to read public idea descriptions and comments before choosing work, but does not require treating that content as untrusted. A malicious submission could contain prompt-injection text that persuades an agent to disclose data, run unsafe actions, or follow unrelated instructions.
The documented workflow explicitly consumes public descriptions and comments before an agent acts, with no prompt-injection boundary or review step. This creates a clear indirect-instruction risk.
Capability review items (2)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
Most command, filesystem, and reconnaissance findings are false positives caused by Markdown backticks, TypeScript template literals, placeholders, and API path documentation. Real risk remains from outbound OpenServ API calls, environment-based credential handling, and get-api-key.ts printing live API and wallet secrets. No prompt injection language was found in the reviewed files.
examples/get-api-key.ts prints the returned API key and generated WALLET_PRIVATE_KEY to stdout for .env setup. Shared terminals, command logs, or CI logs could retain those secrets.
The file directly logs data.apiKey and the generated privateKey. This is a clear secret exposure path when output is captured.
The skill guides agents to submit ideas, pick up work, ship services, upvote, and comment with an API key. These are intentional workflows, but they create account-visible side effects.
The workflow language explicitly tells agents to perform authenticated Ideaboard actions. The risk is business-logic impact, not hidden code execution.
Capability review items (57)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
(process.env.WALLET_PRIVATE_KEY as `0x${string}`) || generatePrivateKey();
The code or runnable documentation handles OPENSERV_API_KEY or WALLET_PRIVATE_KEY. This is legitimate, but live credentials can be exposed or misused if mishandled.
The code or runnable documentation handles OPENSERV_API_KEY or WALLET_PRIVATE_KEY. This is legitimate, but live credentials can be exposed or misused if mishandled.
The code or runnable documentation handles OPENSERV_API_KEY or WALLET_PRIVATE_KEY. This is legitimate, but live credentials can be exposed or misused if mishandled.
The code or runnable documentation handles OPENSERV_API_KEY or WALLET_PRIVATE_KEY. This is legitimate, but live credentials can be exposed or misused if mishandled.
The code or runnable documentation handles OPENSERV_API_KEY or WALLET_PRIVATE_KEY. This is legitimate, but live credentials can be exposed or misused if mishandled.
The code or runnable documentation handles OPENSERV_API_KEY or WALLET_PRIVATE_KEY. This is legitimate, but live credentials can be exposed or misused if mishandled.