issueplanlisted

# issueplan Plan with the current large language model (LLM) in the current session. Keep planning and building in this session. Do not create or use a subagent during this workflow. The current agent owns every step. ## Input Accept a task description and an optional GitHub issue: - Accept prose, such as "issueplan adding X to Y." - Accept a full issue URL, `#<N>`, bare `<N>`, or `owner/repo#N`. - Ask what to plan only when the task is unclear. ## Steps ### 1. Resolve the GitHub issue when one is referenced Fetch the issue before planning: ``` gh issue view <N> --json number,title,body,url ``` Add `-R owner/repo` for another repository. A bare command resolves against the current repository. Stop if the command fails. Do not plan from a paraphrase when the referenced issue is unavailable. Record the issue number, URL, title, and full body. Skip issue posting in step 4 when the user gave no issue. Read any `## Execution` block as a task constraint. Keep the current session model and effort for planning. ### 2. Produce the plan in the current session Do not call the Agent tool, Task tool, workflow delegation, or any subagent mechanism. Inspect the repository with read-only commands. Read its instruction files and the code that controls the requested behavior. Produce a concrete, ordered implementation plan. Include: - Files to create or modify. - The implementation approach and build sequence. - Correctness risks, safety risks, and edge cases. - Tests and oth