create-issuelisted

**Portable helper anchor (single-statement).** This skill invokes helper scripts bundled beside it (`load-prompt-extension.sh` below; `issue-audit-state.py` — the audit-lifecycle state owner — throughout Step 3.6 and Step 4; `resolve-main-root.sh` wherever the canonical draft root is resolved or bound; `ensure-label.sh` / `apply-labels.sh` in Step 4 sub-step 5a). Every call resolves the skill directory **inline at the call site** via `${CLAUDE_SKILL_DIR:-<absolute skill base directory this runner reports in context>}` — the anchor prefers `$CLAUDE_SKILL_DIR` when it is set **and non-empty** (Claude Code; run each command exactly as written, behavior unchanged); otherwise, replace the placeholder with the skill base directory this runner reports in context — e.g. a `Base directory for this skill:` line, observed on Copilot CLI (empty `$CLAUDE_SKILL_DIR` confirmed on Copilot CLI v1.0.67; expected on Cursor, Codex CLI, Gemini CLI, …). The `:-` expansion treats an **empty** `$CLAUDE_SKILL_DIR` exactly like unset because the observed failure on a non-Claude-Code runner is *empty* expansion, not an unset variable. If the reported path is Windows-form (`C:\...`) — which a POSIX shell (WSL bash, Git Bash) cannot use as-is — convert it **before substituting**: run one standalone `wslpath -u '<path>'` (WSL) or `cygpath -u '<path>'` (Git Bash/MSYS2) command and substitute its printed output **only if the command succeeds and prints a non-empty path — otherwise fall through to the drive-