--- name: positioning description: Use when a Guochuang project needs its essence, core advantages, three selling points, innovation framing, market or beneficiary opportunity, historical-method comparison, or track-based gap diagnosis. --- # Project Positioning Apply the shared policies in `gps-common`. When saving a Markdown report, use the common project-name rule with topic `定位`. Turn project facts into a defensible strategic diagnosis. A polished label, historical case, patent, logo, forecast, or performance number is not an owned advantage by itself. ## Owns - Project essence, advantage classification, ranking, and exactly-three selling-point contract. - Innovation framing, opportunity fact/hypothesis separation, rubric-dimension mapping, and historical method transfer. ## Does not own - Intake, eligibility, rubric route, report synthesis, or action assignment (`navigator`). - Source inspection and contradiction resolution (`evidence`); technical validity (`innovation`); customer, pricing, finance, or delivery judgment (`business`). ## Required inputs Use inspected evidence and the selected rubric route. Apply GPS Common conclusion language. Add `evidence` for every positioning pass; add `innovation` for technical claims and `business` for market, customer, price, finance, or social-delivery claims. ## Diagnose 1. State essence as: `target + setting + problem + intervention + outcome hypothesis`. 2. Generate candidates from mechanism, engineering integration, field access, data/feedback, commercial or social-delivery mechanism, and student ownership. 3. Classify each candidate: `owned` (student-attributable and evidence-bounded), `conditional` (promising but condition-limited), `borrowed` (partner/platform/policy/prestige derived), or `pseudo` (publicity, unsupported superlative, source-count, intent, forecast, or honour inflation). 4. Rank by track relevance, differentiation, evidence quality, student ownership, and defensibility. 5. Select exactly three defensible selling points. If fewer exist, show each remaining slot as `missing`; never fill it with a weaker publicity claim. 6. Separate innovation mechanism from its performance result; name the baseline, condition, and limitation before claiming superiority. 7. Separate verified market or beneficiary facts from `hypothesis` opportunity statements and specify the next research task. 8. Read `references/case-patterns.md` only when a historical method pattern fits the unresolved reviewer risk. Transfer its method, never names, wording, figures, or private facts. ## Selling-point record Return one record for each of three slots with these exact fields: `audience`, `high_value_problem`, `distinctive_action`, `conditional_result`, `evidence_ids`, `rubric_dimensions`, `status`, `limitation`. Also return: essence; core-advantage table with classification; innovation versus result distinction; opportunity facts, hypotheses, and research tasks; rubric gap matrix; selected case-pattern IDs and transferable method. Do not calculate an official score or predict an award.