skills/brainstorming/SKILL.md (1)

62-62: Approve with optional wording refinement suggestion.

The guidance correctly establishes the terminal state and skill invocation boundaries. The intent is clear: after design approval, only using-git-worktrees and writing-plans should be invoked, with no other implementation skills allowed.

📝 Optional wording refinement for precision

The phrase "after brainstorming" might be slightly ambiguous since using-git-worktrees is step 7 of the brainstorming checklist. Consider:

-**The terminal state is invoking writing-plans.** Do NOT invoke frontend-design, mcp-builder, or any other implementation skill. The ONLY skills you invoke after brainstorming are using-git-worktrees (for workspace isolation) then writing-plans.
+**The terminal state is invoking writing-plans.** Do NOT invoke frontend-design, mcp-builder, or any other implementation skill. The ONLY skills you invoke from brainstorming are using-git-worktrees (for workspace isolation) then writing-plans.

This clarifies that these skills are invoked as part of the brainstorming workflow, not after it completes.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@skills/brainstorming/SKILL.md` at line 62, Update the wording that currently
reads "The terminal state is invoking writing-plans. Do NOT invoke
frontend-design, mcp-builder, or any other implementation skill. The ONLY skills
you invoke after brainstorming are using-git-worktrees (for workspace isolation)
then writing-plans." to remove ambiguity around "after brainstorming":
explicitly state that within the brainstorming workflow (specifically step 7 of
the brainstorming checklist) the only allowed follow-up skills are
using-git-worktrees and then writing-plans, and that no other implementation
skills (e.g., frontend-design, mcp-builder) should be invoked at that point;
reference the exact phrase "The terminal state is invoking writing-plans." and
the skill names "using-git-worktrees" and "writing-plans" so reviewers can
locate and replace the sentence with a clearer version indicating these skills
are invoked as part of the brainstorming steps rather than only after
completion.

skills/brainstorming/SKILL.md (1)

136-140: Add an explicit “verify worktree active” check before invoking writing-plans.

Right now it requires invoking using-git-worktrees, but doesn’t explicitly require confirming the session is actually in the new worktree before proceeding. Adding that check would further harden the #574 failure mode.

Suggested doc tweak

 **Workspace isolation:**

 - Invoke using-git-worktrees to create an isolated worktree for implementation
 - This ensures implementation work doesn't affect the main working tree
+- Verify you are now operating inside the created worktree/branch before continuing

 **Implementation:**

 - Invoke the writing-plans skill to create a detailed implementation plan
 - Do NOT invoke any implementation skill (frontend-design, mcp-builder, etc.). writing-plans is the next step after workspace isolation.

Also applies to: 144-145

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@skills/brainstorming/SKILL.md` around lines 136 - 140, The documentation
currently instructs invoking using-git-worktrees before running writing-plans
but lacks an explicit verification step that the session is in the new worktree;
update the SKILL.md content around the "Workspace isolation" section to require
and describe an explicit "verify worktree active" check (e.g., run git worktree
list or check GIT_WORK_TREE/GIT_DIR/environment or a provided helper)
immediately after using-git-worktrees and before calling writing-plans so the
flow confirms the active worktree and aborts if not active; reference the
commands/helpers using-git-worktrees and writing-plans in the text so readers
know where to add the check.