issues
Convert existing tasks into actionable, dependency-ordered GitHub issues for the feature based on available design artifacts.
SKILL.mdskills/issues/SKILL.md
GitHubUser Input
You MUST consider the user input before proceeding (if not empty).
Outline
Locate the feature: Call the
akka_sdd_list_specsMCP tool to find features. Verify thattasks.mdexists for the target feature (has_tasksmust be true). Iftasks.mdis missing, instruct the user to run/akka:tasksfirst. Read thetasks.mdcontent from FEATURE_DIR.Build the task-ID set: Scan
tasks.mdfor every task line — each starts with- [ ]followed by a task ID matchingTand three digits (e.g.T001). Collect the set of task IDs you are about to process.Determine the GitHub repository from the git remote:
git config --get remote.origin.url[!CAUTION] ONLY PROCEED TO THE NEXT STEPS IF THE REMOTE IS A GITHUB URL
Fetch existing issues for deduplication: Before creating anything, use the GitHub MCP server's
list_issuestool to find issues that already cover the task IDs from step 2.- Do not pass a
statevalue — omitting it returns both open and closed issues. - Request
perPage: 100to minimize calls. - The tool uses cursor-based pagination: request further pages with the
afterparameter using theendCursorfrom the previous response. - For each issue title, match against the pattern
\bT\d{3}\b(word boundaries soST001andT0010do not false-match). This recognizes titles written asT001 ...,T001: ..., or[T001] .... When it matches one of your task IDs, mark that ID as already having an issue. - Stop paginating as soon as every task ID has been matched, or when there are no more pages. This bounds the number of calls on repos with large issue histories.
- Do not pass a
Create issues for tasks that don't have one yet: For each task in
tasks.md:Strip the leading
- [ ]and any[P]/[US#]markers to recover the task ID and description. Example:- [ ] T001 [P] [US1] Create User model in src/models/user.py→ IDT001, descriptionCreate User model in src/models/user.py.If the ID is in the "already has an issue" set from step 4: skip it and report
T001 already has an issue, skipping.Otherwise, use the GitHub MCP server's
issue_writetool to create a new issue in the repository derived from the git remote. Use the canonical title format:T001: Create User model in src/models/user.py
[!CAUTION] UNDER NO CIRCUMSTANCES EVER CREATE ISSUES IN REPOSITORIES THAT DO NOT MATCH THE REMOTE URL
Report: Summarize the run — total task IDs found, existing issues detected (skipped), new issues created, and the target repository URL.
Done When
- FEATURE_DIR and
tasks.mdwere resolved viaakka_sdd_list_specswithhas_tasks == true. - The full set of task IDs (
T\d{3}matches intasks.md) was collected before any GitHub calls were made. - The git remote was inspected and confirmed to point at a GitHub repository — for any non-GitHub remote the command stopped without creating issues.
-
list_issueswas called withperPage: 100, both open and closed issues (nostatefilter), and cursor-based pagination viaafter; pagination stopped as soon as every task ID was matched or the pages ran out. - Titles were matched against
\bT\d{3}\bwith word boundaries, so tokens likeST001orT0010were not false-matched. - Every new issue used the canonical title format
T###: <description>, built by stripping- [ ],[P], and[US#]markers from the source task line. - Tasks whose IDs already had an existing issue were skipped and each skip was reported (e.g.
T001 already has an issue, skipping). - Every created issue lives in the repository derived from the git remote and in no other repository.
- The report includes the total task count, count of skipped-because-already-exists, count of newly created issues, and the target repository URL.