extend

Sub-idea: $ARGUMENTS

If $ARGUMENTS contains a ticket reference, invoke swe-workbench:ticket-context first and prepend its structured summary to the delegation context below. Skip if $ARGUMENTS is free-text with no recognizable ref. (Trigger patterns are defined in that skill's "When to invoke" section.)

PR detection (inline):

PR_JSON=$(gh pr view --json number,state,headRefName,isDraft,url 2>/dev/null || true)
PR_STATE=$(printf '%s' "$PR_JSON" | python3 -c "import sys,json; d=json.load(sys.stdin); print(d.get('state',''))" 2>/dev/null || true)
PR_NUM=$(printf '%s' "$PR_JSON"   | python3 -c "import sys,json; d=json.load(sys.stdin); print(d.get('number',''))" 2>/dev/null || true)
HEAD_REF=$(printf '%s' "$PR_JSON" | python3 -c "import sys,json; d=json.load(sys.stdin); print(d.get('headRefName',''))" 2>/dev/null || true)
PR_URL=$(printf '%s' "$PR_JSON"   | python3 -c "import sys,json; d=json.load(sys.stdin); print(d.get('url',''))" 2>/dev/null || true)
IS_DRAFT=$(printf '%s' "$PR_JSON" | python3 -c "import sys,json; d=json.load(sys.stdin); print('true' if d.get('isDraft') else 'false')" 2>/dev/null || true)

Note: gh pr view resolves against origin; on fork-based workflows where origin is the fork, it may return no PR. If that happens, use gh pr view --repo <upstream-owner>/<repo> explicitly.

If PR_STATE is "OPEN": capture PR_NUM, HEAD_REF, PR_URL, and IS_DRAFT into context. If IS_DRAFT is "true", note "Draft PR detected — it will not be auto-marked ready-for-review." Before activating swe-workbench:workflow-extend, complete the following interrogation step (OPEN branch only):

Interrogation mode. Before producing anything, resolve the mode:

• Explicit signal in the invocation is honored without asking. grill-me = --grill, "grill me", or "grill-me mode". standard = --standard, "standard", or "quick". Strip the signal from $ARGUMENTS and record the resolved mode.
• No explicit signal: ask via AskUserQuestion — one question, header "Mode", options Standard (recommended, listed first) and Grill me. Standard description: "Lightweight clarify — a restatement and at most one question, then proceed." Grill-me description: "Relentlessly walk the decision tree one question at a time, each with a recommended answer, self-answering from the codebase where possible." Use the user's choice.

Standard mode: proceed with the command's existing lightweight clarify (a restatement and at most one clarifying question) — do not ask the mode question again.

Grill-me mode: activate swe-workbench:workflow-grill and run its interrogation loop to completion (exit on shared understanding or when the user says "proceed"). Then thread the emitted ## Resolved decisions block into the command's normal artifact/delegation step below — the same way a ticket-context summary is prepended — and continue as in standard mode.

Then activate swe-workbench:workflow-extend with all four values.

The list below is a visible contract of the phases swe-workbench:workflow-extend runs — it does not replace the skill activation above. Execute phases 2–5 in order. Phase 1 (Branch) is intentionally skipped — the open PR branch is reused, so no new branch or worktree is created.

Phase 2 — Implement Execute the plan via superpowers:executing-plans or superpowers:subagent-driven-development. Apply swe-workbench:principle-tdd per unit: red → green → refactor.

Phase 3 — Verify Run superpowers:verification-before-completion before claiming any phase done. Do not advance to Phase 4 until format / lint / test pass with evidence.

Phase 4 — Review Dispatch BOTH reviewers IN PARALLEL — in a single batch (same turn), as two distinct required invocations, neither optional:

• superpowers:requesting-code-review (a Skill) — plan-alignment, standards
• swe-workbench:reviewer (a subagent) — diff correctness/security/design in Severity | File:Line | Issue | Why it matters | Suggested fix format