Stop deciding by gut feel.

Per the master prompt: (1) incumbent status — is there one, who, and what is their relationship to the agency; (2) capability match — do your NAICS, certifications, and past performance align with the SOW; (3) past performance fit — have you done similar work, same agency, same NAICS; (4) teaming need — can you prime, or is this a teaming play; (5) price-to-win signal — any public indicator of expected value or competitive range; (6) agency relationship strength — prior dialogue with the contracting officer or agency; (7) SOW clarity — is the SOW well-scoped or ambiguous and mid-pivot; (8) schedule feasibility — given the proposal due date, can you realistically write a competitive response.

0–5 per dimension gives a max of 40 total. The thresholds (≥25, 16–24, <16) split that range into three actionable bands with enough resolution to distinguish a marginal Go from a strong one. A score of 0 is valid — it means no signal or an explicit negative on that dimension. Leaving a dimension blank also results in 0, which is intentional: a blank is a data gap, and data gaps hurt the score to reflect real uncertainty.

≥25/40 (62.5%) means most dimensions have positive signal — a credible prime bid. 16–24 means the pursuit has real gaps that teaming can potentially address: weak past performance, thin agency relationship, or a teaming need for a specialized sub. Below 16/40 the gaps are structural, not patchable by teaming — the recommendation is No-Go. [VERIFY] — these thresholds are calibrated to the sample set used during pack development; apply judgment if your pursuit profile is unusual.

It means the pursuit is viable with a specific teaming action — identify a partner who covers the gap the matrix surfaced. The matrix will name the gap (past performance fit, specialized technical capability, incumbent relationship) in the top three reasons. Your next step is a teaming conversation, not a proposal kickoff.

Not through the pack as shipped. The 8 dimensions are fixed in PROMPT.md — the prompt is designed around this specific set, and the threshold math assumes all 8 are scored on the same 0–5 scale. If you need a different dimension set, you'd modify PROMPT.md directly, which you can do since it's a plain text file you own. That's outside the pack's supported use case, but the file is yours.

The prompt refuses to score. It returns a list of the blank dimensions and stops. This is intentional — a matrix with 4+ blanks produces a recommendation with too little signal to act on. Fill the missing dimensions (even partial notes are better than nothing) and re-run.

14-day no-questions refund. Open the zip, run it on one real pursuit, decide. If it doesn't change how you make that decision, refund it.

It's standalone. You paste your filled INPUT-TEMPLATE.md into a Claude.ai chat alongside the master prompt. Output is markdown — copy it into your CRM note, your pursuit file, or your team channel. There's no API, no webhook, no native integration with GovWin, Salesforce, or your BD pipeline tool. That's by design at this price point.