Open Issues Need Help
View All on GitHubAI Summary: This issue proposes an enhancement to an automatic assignment feature. The system should periodically search for a specific button every second, and upon discovery, introduce a one-second delay before clicking it. The primary goal is to simulate human interaction timing, making the automated process feel more natural.
AI Summary: This issue requests an enhancement to the 'search splash window' that appears during an auto-assignment process. The current window should be updated to also display the real-time status of the ongoing search, providing users with more information about the operation's progress.
AI Summary: The issue reports that an automatic assignment process failed, and critically, this failure led to the entire application crashing. This indicates a severe bug where an expected or recoverable failure path results in an unhandled exception or critical system instability.
AI Summary: This issue requests that the application's splash screen color should adapt to the operating system's dark mode setting. This would involve detecting the current OS theme and dynamically applying an appropriate color scheme to the splash screen upon launch.
AI Summary: This issue aims to improve user experience by adding visual cues that allow users to clearly distinguish between two specific browser opening behaviors: one where the browser opens and automatically clicks a button, and another where the browser opens during "phase 3." The goal is to make these two scenarios visually identifiable.
AI Summary: This issue requests a UI enhancement to improve the visibility and clickability of URLs. The proposed solution is to apply a distinct color to URLs, making them easier for users to identify and interact with.
AI Summary: This issue aims to fix a bug related to the 'eyes' condition used for determining the LLM's operational state. To resolve this, the system will no longer rely on 'eyes' in its conditional logic. Instead, it will reference the latest entry in `llm_statuses` (the most recent LLM status) to accurately determine if the LLM is in a 'Working' state or 'phase3'.
AI Summary: The automatic identification for three types of coding agents has passed tests. As a follow-up, a PR author display feature needs to be made configurable. It should be toggleable via TOML, with the default setting being 'off'.
AI Summary: The user reports that a feature is unexpectedly performing a dry run despite being configured for full execution. Specifically, the `enable_execution_pr_title_fix_comment` setting in the TOML file is set to `true`, which should enable the actual execution of a PR title fix comment. However, the system is still only performing a dry run.
AI Summary: This issue reports that raw JSON data is being directly written into markdown files stored within the `pr_phase_snapshots/` directory. This likely results in unformatted and difficult-to-read content in these markdown files, indicating a problem with how data is being embedded or rendered during the file generation process.
AI Summary: The LLM Working judgment sometimes fails due to insufficient data. To resolve this, the task is to use `curl` to fetch and save the HTML of the relevant page. Additionally, this saved HTML should be converted into a Markdown format and also saved, with all files stored under the `pr_phase_snapshots/` directory.
AI Summary: The Codex Coding Agent occasionally generates Pull Requests with the generic and potentially inappropriate title "Addressing PR comments." This issue occurs intermittently, suggesting the agent might be misinterpreting its task or defaulting to a placeholder title when a more specific one is expected.
AI Summary: This issue requests an update to the Japanese README file (README.ja.md). The documentation needs to be revised to accurately reflect the current state of the project's implementation and any changes to its default settings.