nextlisted

# Next: Issue Triage & Recommendation Analyze open GitHub issues, build a dependency graph, and recommend the highest-impact issue to work on next. ## Workflow ```text 1. Fetch & Pre-process → 2. Build Dependency Graph → 3. Detect Cycles → 4. Rank Issues → 5. Present Recommendations → 6. Select & Assign ``` ### Step 1: Fetch and Pre-process Issues Run the `open-issues.sh` script from the plugin's `scripts/` directory: ```bash # Auto-detect repository from current directory sh "${CLAUDE_SKILL_DIR}/../../scripts/open-issues.sh" # Or specify repository explicitly sh "${CLAUDE_SKILL_DIR}/../../scripts/open-issues.sh" OWNER/REPO # Include assigned issues in results sh "${CLAUDE_SKILL_DIR}/../../scripts/open-issues.sh" --include-assigned ``` **Validate the response:** - Check stderr for a JSON `error` object — if present, abort with the error message (auth failures, missing tools, repo detection failures are handled by the script) - Extract and store the full JSON output as `$ISSUES_DATA` **Edge case — no issues:** If `.total_open == 0`, report "No open issues found" (if `--include-assigned` was used) or "No unassigned issues found (try `--include-assigned` to include assigned items)" (default mode) and stop. **Edge case — 100+ issues:** If `.total_open >= 100`, warn: "Results capped at 100 issues. Consider filtering by label or milestone for a broader view." The JSON output contains pre-computed fields for each issue: - `blocked_by`, `blocks` — parsed from issue body p