coreyhaines31/makerskills / skills/product-manager
product-manager
VerifiedWhen you want to plan, build, or check an app from idea through launch readiness — web, Apple, mobile, or extensions. Writes the brief, picks the stack (Next.js + TypeScript + Prisma by default, Swift on Apple), designs complete UX states, defines roles and agent access, measures performance before optimizing, and produces a readiness report where missing evidence is never a pass. Four modes: spec (plan v1), check (launch readiness), audit (ranked improvements, including what a new model changes), fleet (a named list of apps). Triggers on "/product-manager," "act as my product manager," "PM this app," "build me an app," "plan v1," "spec this app," "is this app ready to launch," "launch readiness," "audit my apps," "what's missing before we ship," "do a comprehensive audit," "how could we improve my app," "a new model just came out." Not for launch marketing or growth (see launch), integration plumbing (see toolify), CI cost (see ci-audit), or managing a project backlog/kanban (see project-manager).
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 16.5 KB |
| references/audit.md | markdown | 11.8 KB |
| references/defensibility.md | markdown | 5.6 KB |
| references/performance.md | markdown | 5.4 KB |
| references/readiness.md | markdown | 6.2 KB |
| references/standards.md | markdown | 7.0 KB |
| references/trust.md | markdown | 11.0 KB |
| references/ux.md | markdown | 9.7 KB |
references/readiness.md
View on GitHub# Testing, launch readiness and the report
## Testing
Map each critical journey and invariant to the cheapest meaningful proof:
- **Unit tests** for policy and pure transformations.
- **Integration tests** for DB queries, tenant boundaries, webhooks and idempotency.
- **E2E tests** for a few real value journeys, including keyboard and recovery paths.
- **Visual checks** at representative sizes catch overflow. Screenshots don't prove interactions or authorization.
Use synthetic or scrubbed fixtures. Before any test that could send mail, charge a card or publish, confirm the service is faked or sandboxed. A staging environment can still email real people.
CI, as the profile warrants: typecheck, lint, tests, build, dependency and secret scans, and a migration compatibility review. Don't demand a web build from a native utility. Migrations need expand/contract compatibility where relevant, plus a rollback or forward-fix plan. For existing GitHub Actions cost and timing, use **ci-audit**.
Before calling it ready, exercise these failure modes: denied actor, revoked access, double-submit, network timeout after a side effect, partial upload, stale cache, backend outage, recovery after interruption, the key accessibility journey, and (if the app ships light and dark) switching Light / Dark / System and reloading. Anything you didn't exercise is *not assessed*.
## Readiness inventory
- **Core journeys:** success plus the important denial, error and recovery cases.
- **Trust:** authorization and tenant isolation, secrets, sessions, input and uploads, consent, privacy, data lifecycle. Unresolved risks have an owner.
- **Billing / AI** (if applicable): entitlements, limits, signed idempotent webhooks, failure handling.
- **Performance:** representative baselines and bottlenecks, not one universal score.
- **Delivery:** reproducible build, config inventory (no secret values), environment separation, migration review, release approver.
- **Recovery:** a rollback or forward-fix plan, and backup schedule plus **an actual restore test**. A "backup succeeded" flag isn't restore proof.
- **Operations:** actionable monitoring, privacy-safe logs, alert owner, cost guardrails, an incident runbook, degraded-mode behavior.
- **Docs:** setup, use, update and uninstall; repo invariants and test commands for coding agents; security reporting and support contacts.
- **Platform:** signing and notarization, mobile permissions and store requirements, extension permissions and updates. Actually submitting is a separate step.
## Status vocabulary
| Status | Meaning |
|---|---|
| **pass** | Verified at a stated revision and environment with enough evidence |
| **gap** | Verified as unmet. Say whether it blocks |
| **not assessed** | Unknown, no access, not enough evidence, or not tested |
| **justified N/A** | Doesn't apply to this profile, with the reason |
- Missing evidence isn't a pass.
- Searching and not finding something is an observation, not a verified gap. But reading the whole relevant code path and confirming the mechanism isn't there (e.g. a webhook handler with no event-ID dedupe and no table to store processed events) *is* a verified gap.
- Reading code can't make a mechanism that's present a pass. You have to exercise it.
- A named manual run (who, when, which revision and environment) is a pass only for the paths actually exercised. The denial and recovery paths around them stay *not assessed*.
- A *not assessed* row on a critical control (tenant isolation, secrets, restore you rely on, a release path that has never run) is marked "not assessed, blocking". On its own it makes the verdict **unproven**. A verified blocking gap anywhere makes it **not ready**.
- An approved exception on a blocking gap doesn't change the verdict. The exception is different when its control removes the exposure, e.g. the feature is flagged off for launch. Then the row becomes a non-blocking gap.
- Absence of telemetry, confirmed by searching dependencies and network code, supports "justified N/A". Say it was a search.
- An approved exception keeps the item as a gap. Record the owner, compensating control and review date.
- No single numeric score.
- Keep deployment stages distinct: drafted, checked, merged, deployed and live-verified are different states.
## Report template
```markdown
# Readiness — <project> (<profile>, <mode>)
Scope: <revision / environment / what was and wasn't accessible> · <date>
Verdict: <not ready | ready for review | unproven> for this scope
| Requirement | Status | Evidence / limitation | Owner → next step |
|---|---|---|---|
## Blockers
<observed behavior, journey, evidence, risk, the test that proves the fix>
## Unknowns
<missing access or coverage → smallest safe next check>
## Next step
<one concrete action, and what approval it needs>
```
### Worked example (synthetic, illustrative results)
Scope: a collaborative web SaaS staging candidate at ref-17, checked from docs and isolated test artifacts with no production access. Verdict: **not ready for release approval.**
| Requirement | Status | Evidence / limitation | Owner → next step |
|---|---|---|---|
| O1 tenant reads | pass | A/B read/write + removed-member tests passed at ref-17 | Eng. No production claim |
| O2 video feedback | gap | Screenshot works. Video exceeds the supported type list | Product: build it or record an exception |
| O3 private downloads | **gap, blocking** | Signed route served another audience's attachment | Security: block release |
| O4 event privacy | not assessed | Only the event plan doc was available | Analytics: sandbox payload check |
| O5 agent write retry | pass | Repeated draft creation returns the same result | Eng |
| O5 agent feedback tool | gap | MCP has no `submit_feedback`; agent reports of missing tools go nowhere | Eng: add it, routed to the triage inbox |
| Performance: summary read | not assessed | One cold lab run, so not comparable | Perf: bounded comparable trials |
| Restore | not assessed, blocking | Backup-success flag only | Ops: isolated restore drill |
| Extension permissions | justified N/A | No extension ships | — |
Next step: fix attachment isolation (needs implementation approval). The restore drill comes after that.
Ready-for-review isn't permission to deploy, publish or announce. Launch posts and growth work go to the marketing skills.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |