agentsclimarketplace

Go nogo memo

Skill stanislavnianko/product-discovery-claude-skills/plugins/discovery-phase/skills/go-nogo-memo

Curated Claude skill pack for structured product discovery

Install
npx -y skills add stanislavnianko/product-discovery-claude-skills --skill go-nogo-memo

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

[discovery-phase pack · deliverables] Produces a single GO / PIVOT / KILL memo consolidating all evidence into a one-page decision document. The only artifact most senior stakeholders read. Adapted for outsourcing — decision-owner is typically the client, agency provides recommendation with evidence chain.

SKILL.md

4.8 KB, as published. Nobody here has run it

Go / No-Go Memo

Part of the discovery-phase skill pack · deliverables group · reads discovery-context.md (run profile-builder first if missing).

The terminal decision document. Everything before this fed into it; everything after follows from it. In outsourcing, the agency recommends, the client decides — the memo must support both perspectives.

When to run

  • End of any discovery cycle
  • Standalone if BA brings in evidence from elsewhere and just needs to write the memo
  • Re-run after major new evidence (re-versioned)

Differs from proposal: proposal is commercial (here's what we'd do, please sign); memo is decisional (here's what we believe, please decide). They can co-exist.

Step 1 — Read context + scan evidence

Read discovery-context.md (section 5. Expected deliverable). Then scan ./discovery/ for available artifacts — even partial discovery can produce a memo, but flag the gaps explicitly in the memo's "Evidence" section.

If discovery-context.md is missing, ask the BA inline: "who's the decision-owner (client / agency / both) and what decision are we framing (proceed / pivot / kill)?" — tag the output [ASSUMED FRAMING]. Never block; recommend profile-builder for high-stakes work.

Step 2 — Force the decision first

Before writing anything, ask the BA:

"Based on what we have, what's the recommendation: GO / PIVOT / KILL? One word. We'll justify on paper next."

If they hesitate, probe:

  • "What's the one finding that would flip this to the opposite answer?"
  • "If you had to bet 10% of the agency's quarterly capacity (or your own time) on this, would you?"
  • "Imagine this memo arrived from another team. Would you approve their recommendation?"

The memo serves the recommendation. No defensive prose.

Step 3 — Clarify the decision split (outsourcing)

Two roles must be explicit:

  • Recommendation owner — the agency BA / lead. Owns the recommendation in the memo.
  • Decision owner — typically the client (per discovery-context.md engagement). The memo gives them what they need to decide.

State both at the top of the memo: "We (agency) recommend X. Decision rests with <named client decision-maker>."

Step 4 — Fill the template

Fill ./template.md. It owns the structure (recommendation line 1, TL;DR, rationale, evidence chain, key findings, risks-remaining, decision-specific GO/PIVOT/KILL section, signatures). This skill's job is the content — the template is the shape.

Key rules while filling:

  • Line 1 is the recommendation. Not buried, not in paragraph 3.
  • Evidence chain links every artifact in ./discovery/. A reader who disagrees can audit in 20 minutes. Mark missing artifacts explicitly (e.g., ~~Tech Spike Report~~ — not run).
  • Key findings = one-sentence claim + one quote / data point each. Quotable in future planning docs.
  • Risks-remain comes from risk-assumption-map.md rows NOT retired during discovery, with owner per row. Distinguish client-owned (post-decision) vs agency-owned (if engagement converts).
  • Decision-specific section — fill ONE (GO / PIVOT / KILL), delete the others.
  • Client decision signature — leave blank until the client decides.

Step 5 — Output + index update

./discovery/go-no-go-memo.md per ./template.md. Update DISCOVERY.md (or _log.md) index entry to: "Cycle complete — recommendation: <X>; decision: pending/<Y>".

Append to _log.md: [go-nogo-memo | YYYY-MM-DD] recommendation: <GO/PIVOT/KILL>; decision_owner: <client/agency>; commercial_path: <proposal/sow/handoff/none>.

Anti-patterns

  • "Needs more research." Almost always KILL or PIVOT in disguise. Name it. If you genuinely need another cycle, name the specific question, the specific test, and the cost — otherwise "more research" is a graveyard.
  • GO by default. Teams hate killing projects they researched. Apply the third-party test: "would you approve this if it came from another team?"
  • Buried recommendation. Line 1, no exceptions.
  • Memo without evidence links. Claims rot; linked claims survive.
  • Conflating recommendation and decision. In outsourcing, agency recommends, client decides. Be explicit.
  • No risks-remain section. This is where the engagement quietly fails post-launch. Capture them with owners.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.