RizwanZafaris
UserRun a product from discovery to sunset: gated templates, PM canon cards, and AI skills that work without AI.
Categories
Indexed Skills (24)
tutor
Teaching-posture reuse of the Conductor's question machinery for deliberate practice on fictional products. Use when a user says "teach me", "quiz me", "tutor", "drill me", or asks to practice a stage, a template, or a step from a learn/ path. The learner fills a template for a fictional product; the tutor critiques against the template's exit gate citing knowledge cards, asks from the real stage banks, pushes once with the challenge grammar, then shows a model answer and scores 0, 1, or 2. Reads the Conductor's banks and protocol strictly read-only and never touches a real product workspace.
ai-prd
Write and pressure-test a PRD for an AI or agentic system. Use when a product manager is specifying a feature where the implementer is a model, not an engineer - LLM features, agents, copilots, ML-backed decisions - or when reviewing an existing PRD that says "the AI should" anywhere. Turns acceptance criteria into eval sets, makes non-determinism an explicit requirement, and treats guardrails as features with owners. For products under a financial or data regulator, routes to the regulated module and its byte-exact template.
competitive-intel
Produce a sourced competitor teardown (offer, pricing, positioning, gaps) where every claim carries a URL and a date, landing in the competitive analysis and positioning templates. Use when a build, buy, pricing, or positioning decision needs to know what the alternatives actually do, when a competitor move has the team reacting on rumor, or when win-loss keeps naming the same rival. Takes the decision, the job, and a candidate set; returns the evidence table, a teardown per competitor, the structural read, and a committed so-what.
conductor
Interview-driven operation of the six-stage product loop. Use when a user says "start" on a new product, "resume" or "where are we" on an existing one, or asks to be walked through a stage rather than handed a template. Asks one question at a time from the stage's question bank, cross-examines weak answers with a capped challenge grammar, lands every accepted answer in STATE.md and the target template, renders the stage gate on evidence, and never advances past a gate silently.
escalation
Turn a stuck decision or a blocked dependency into a one-page brief (Situation, Impact, Urgency, Options, Recommendation, Ask) and route it up a named ladder with SLAs. Use when a decision has missed its needed-by date, when two teams deadlock, when a risk owner will not act, or when a governance forum keeps noting things without deciding anything.
experiment-designer
Turn an assumption or a growth bet into an experiment brief with a mechanism-stated hypothesis, one primary metric, guardrails with floors, an exposure design, a sample size reasoned from the minimum detectable effect, and stop rules written before launch. Use when a growth plan bet needs a test, when an assumption register row is low confidence and high impact, when a price change should be tried on a slice first, or when someone says "let us just A/B it". Takes the assumption, the metric candidate with its baseline, the eligible traffic, and the guardrail candidates; returns the test card, the filled brief, the sizing record with its calculator inputs, and the pre-committed decision rule.
feedback-synthesis
Turn a pile of raw customer feedback into evidence-weighted themes that land in the discovery templates. Use when a PM has interview transcripts, support tickets, sales-call notes, app reviews, or survey free text and needs themes with counts, contradictions, and a so-what, rather than a word cloud or a quote reel.
gtm-launch-planner
Tier a launch, choose the first cohort, build the channel plan, and draft the comms and readiness lines so Gate 5 finds them already written. Use when a customer-facing release is coming and the GTM plan or the comms table is empty, when PM and PMM disagree on how big the launch is, when a release takes something away or changes what customers pay, or when a failed launch needs a second attempt. Takes the positioning, the success signal, the release scope and window, and the channel evidence; returns the tier decision, the filled GTM plan, the channel plan by funnel stage, the comms plan, and the readiness rows.
launch-readiness
Run the Gate 5 release readiness checklist item by item against the delivery documents and the release candidate, and produce a go, no-go, or conditional go with every condition named, owned, and dated. Use when a release is inside two weeks of its target date, when a readiness document is being filled for the first time, when a launch slipped and the checklist needs re-running, or when someone wants to ship on a feeling. Takes the release readiness document, UAT results, rollback evidence, comms drafts, and the runbook; returns the item-by-item verdict table and the decision packet for the gate chair.
market-sizing
Size a market top-down and bottom-up as TAM, SAM, and SOM, reconcile the two within a stated tolerance, and register every input as an assumption with a source or a test. Use when a business case, a where-to-play choice, a board question, or a Gate 1 cost-of-inaction calculation needs a market number that survives the question "where did that come from". Takes a segment definition, a unit, a price with its source, and count sources; returns both sizings, the reconciliation, a sensitivity range, and the assumption rows.
metrics-tree
Build the chain from a north star metric to its input metrics, to one definition per metric, to a dashboard specification someone can read in a minute. Use when a product has no agreed north star or the one it has fails the vanity test, when two teams report the same metric with different numbers, before OKRs or an instrumentation spec are written, or when a dashboard has decayed into a wall of tiles. Takes the value statement, what is instrumented, the analytics platform, and the audiences; returns the north star sheet, the input tree, the metrics dictionary rows, and the dashboard spec.
okr-critic
Draft or critique an OKR set so that every key result is an outcome someone could score without asking what the team did. Use when a team is writing OKRs for a period, when a draft needs review before signature, when the key result column has filled with tasks and launches, or when team and company OKRs do not add up. Takes the draft set, or the strategy and last period's scores; returns a critique table with a verdict per line, rewritten key results with baselines and owners, and a cascade check, in the OKR template.
pricing-packaging
Design the willingness-to-pay research, the tiers, and the migration rules for a price or packaging change, and land the decision in the pricing and packaging template. Use when a new product or tier needs its first price, when features move between tiers or the value metric changes, when win-loss keeps naming price, or when a price change is about to reach existing customers. Takes the positioning, the value metric candidate, billing and usage data, and dated competitor prices; returns a research design, findings with their caveats, a tier table, migration rules, and the decision log entry.
product-analyst
Run a disciplined research pass as a single analyst with one chat model. Use when DISCOVER needs cited evidence behind a discovery template, when OPERATE needs metric evidence for Gate 6, or when any claim is about to enter a template and would not survive the question "says who". Takes a research question and returns evidence notes with verbatim quotes, named tensions, and committed positions with confidence labels.
product-review
Run the weekly product review as a truth-seeking WIP walk instead of a pitch meeting. Use when a product org needs a standing review ritual, when existing reviews have decayed into status theater or demo days, or when leadership keeps discovering work at launch that they should have shaped at the idea stage. Enforces the 48-hour pre-read, the per-team 20-minute walk across pre-build, in-progress, and post-build work, and decisions that land in the decision log the same day.
program-premortem
Run a premortem on a program plan against the twelve ways program managers actually fail. Use when a program is approaching kickoff, a major cutover, or Gate 3, or when status "feels fine" but nobody can prove it. Takes a plan, status pack, or verbal description and returns which failure modes are already present, with the evidence and the smallest intervention for each.
reg-gap-check
Regulatory gap analysis for payments and fintech product specs. Use when a PM has a PRD, feature spec, or roadmap item touching money movement, cards, wallets, lending, crypto or customer data, and needs to know what a regulator, scheme, or compliance reviewer will challenge before building. Outputs a severity-ranked gap table with owners and evidence needed. Routes verified regulatory citations through the regulated module and never invents regulator text.
roadmap-builder
Convert a backlog and a strategy statement into a scored, sequenced, defensible roadmap. Use when a PM has a pile of asks (features, integrations, compliance items, tech debt) and needs a prioritized quarter-by-quarter plan that survives an executive review - RICE scoring with domain-aware guidance, launch gates, dependency and risk columns, and a regulatory calendar overlay.
stakeholder-update
Write the executive or board update as a situation, complication, resolution narrative that puts the decisions needed on the first page, with every number traced to the metrics review. Use when a QBR or board update is due, when a monthly executive review needs a pre-read, when a one-page update must get a decision this week, or when a sponsor asks "where are we" and the honest answer is longer than a sentence. Takes the filled metrics review, the OKR sheet, the decisions needed with their options, the risk register, and the audience; returns the headline, the decisions table, the metrics and risks sections, the next-period slice, and the objection sheet for the presenter.
story-writer
Turn a signed PRD into epics, user stories, and acceptance criteria that a tester can run and a team can slice into releases. Use when a PRD or one-pager is approved and engineering needs a backlog, when stories in the tracker have no acceptance criteria, when a release plan cannot say which slice ships first, or when a story is too big to estimate. Takes the PRD, the personas, and the FRD where one exists; returns a story map, INVEST-checked stories with IDs, and given/when/then criteria in the acceptance criteria template.
strategy-critic
Stress-test a product strategy with two explicit checks, Rumelt's kernel test and the Playing-to-Win "what would have to be true" test, plus a durability check against the seven powers. Use when a strategy draft needs review before signature, when leadership suspects it is a goals slide, when the roadmap contains bets the strategy cannot explain, or before annual planning copies last year's strategy forward. Takes the strategy document and its evidence links; returns a verdict per check, the conditions that would have to be true with their evidence status, and the rewrites needed, in the product strategy template.
user-interview
Build an interview guide from a research question, run the session on a note sheet that keeps facts apart from compliments, and turn each session into an evidence note. Use when a discovery question needs answers from real users, when Gate 1 cites fewer than five conversations, or when a persona or job statement is still marked assumption. Takes the research questions, a segment, and the team's current beliefs; returns the guide, one note sheet and one evidence note per session, and a handoff to feedback synthesis.
write-prd
Write the general-purpose PRD, and the requirements stack around it (one-pager, BRD, PRD, FRD, NFR, business rules, PR FAQ), sized first against the weight ladder in os/WHICH-DOCUMENT.md. Use when a PM needs to write a PRD or any document in that stack for a feature a human engineering team will build, when a draft has a features table with no objective above it or targets nobody agreed to, or when a request for a PRD has nowhere else in this repository to land. Routes away immediately, before drafting continues, when the implementer is a model (to ai-prd) or a financial or data regulator is in scope (to reg-gap-check). Takes discovery evidence and a weight decision; returns the filled document with kill criteria and a clean pass from spec-review.
lifecycle-conductor
Resume a product journey and report its next gate-aware step. Use when a product needs bounded lifecycle coordination or resume support.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.