Command
Use my supplied visible host, target audiences, competitor list, and days of 7, 30, or 90. Confirm access using list_projects, then read list_prompts({host,limit:100}) and follow nextCursor via cursor. Read get_overview({host,days}) for measurement context and inspect selected get_prompt_runs({host,prompt_id,days}). Audit brand-leading wording, competitor exclusions, duplicated intent, and uneven buyer-stage coverage against my supplied priorities. Separate subjective audit judgments from returned metrics; do not assign an invented bias score. Propose neutral rewrites and a balanced candidate set, explaining tradeoffs for longitudinal comparisons. Return a reviewable change plan only; these tools cannot edit or create prompts.What you need
- Supply a visible host, target audiences, competitors, and a supported window.
- Apply approved prompt revisions manually and preserve the old inventory for comparison.
- The external reference skill targets another provider; use the Kept prompt above or adapt its tool calls before running it.
Expose leading wording and coverage imbalance
Review the actual prompt inventory for phrasing that presupposes the brand is best or excludes relevant alternatives. Pair these editorial judgments with audience coverage checks, making clear which observations come from project data and which depend on your supplied strategy.
Rewrite neutrally without erasing history
Offer neutral alternatives and a proposed balance across intents, audiences, and competitor contexts. The change plan explains how revising the prompt set may affect future comparisons, so reviewers can preserve a baseline and approve deliberate changes before manually updating their project.