The Agent Skills Directory

[SAFE]: The skill provides a robust and logical methodology for software debugging, with clear instructions for environment verification and structured reporting. It does not attempt to bypass any AI constraints or safety protocols.
[COMMAND_EXECUTION]: The workflow suggests running standard development commands like pnpm install and pnpm dev specifically for local reproduction of issues. These are legitimate activities within the intended scope of a debugging skill.
[EXTERNAL_DOWNLOADS]: The skill references standard automation frameworks (Playwright) and specialized tools (agent-wallet) as the primary engines for browser-based testing. It does not contain commands to fetch or execute untrusted remote scripts.
[DATA_EXFILTRATION]: Although the skill involves Web3 wallet operations (e.g., using private keys via the agent-wallet tool), these operations are described as part of a local reproduction sequence for dApps. No instructions were found to exfiltrate credentials or sensitive project data to external domains.
[PROMPT_INJECTION]: The skill's instructions are focused on guiding the agent through a technical workflow. No malicious patterns such as 'ignore previous instructions' or attempts to extract system prompts were detected.

debug-workflow