Launch Window Monitor
Track launch-channel signals, KPI movement, and anomalies from preflight through T+30.
The skill limits itself to WebFetch, treats fetched and pasted content as untrusted, asks for confirmation before saving, and submits registry facts through a propose path. The supplied security policy also documents non-persistent keys, bounded connector behavior, and dry-run defaults for mutations. Five points are deducted because data flows for platform fetches, analytics exports, and runtime memory remain dependent on host configuration; rollback and least-privilege details are not fully specified in this skill, and Product Hunt commercial-use approval and attribution requirements remain an operational risk.
The instructions define manual fallback when connectors or keys are missing, conservative polling cadence, stopping when targets are absent, abnormal-input labeling, and handoff conditions, with no obvious internal contradiction. Twelve points are deducted because the supplied files do not show skill-specific tests, reproduced connector behavior, or controlled handling of actual abnormal API responses; under the static calibration this score cannot exceed 10.
Triggers, time window, inputs, outputs, and non-fit skills are clearly described, with English and Chinese labels and a manual-paste fallback. Four points are deducted because core monitoring surfaces and news retrieval depend on overseas services whose mainland-China reachability is not demonstrated; market, platform-variant, and incomplete-analytics boundaries are also under-specified.
The skill provides frontmatter, version, license, Quick Start, contract, data sources, instructions, save guidance, references, next-skill routing, and termination rules, giving it strong information architecture. Five points are deducted because the supplied material does not establish a skill-specific changelog, maintenance owner, or clear update path; publisher provenance is unverified, and SECURITY.md still lists 17.x as supported while the skill declares 18.0.0.
Expected outputs, measurement labels, threshold alerts, attribution rules, spike-versus-sustain analysis, owned-capture analysis, and handoffs are concrete enough to support a usable report. Nine points are deducted because connectors and actual outputs are not statically verified, while the workflow depends on user targets, analytics exports, and manual platform data; static calibration limits this dimension to 7.
The skill separates Measured, User-provided, and Estimated values and explicitly identifies the flamewar ratio as a heuristic with stated source character; the repository also supplies general CI and test materials. Six points are deducted because there is no skill-specific execution evidence, representative output, or independent multi-source reproduction, and static calibration prevents a score above 5.
- Do not treat the HN comments-to-points ratio as a platform rule or standalone decision signal; the skill labels it as an Estimated heuristic.
- Verify mainland-China reachability for HN, Product Hunt, App Store, and GDELT before relying on live polling; otherwise use pasted values and retain the User-provided label.
- Before using the Product Hunt connector, confirm token, commercial-use approval, and attribution requirements; platform self-reported metrics must not replace owned analytics as attribution truth.
- Confirm that analytics exports and runtime memory contain no credentials, unnecessary personal data, or other information that should not be persisted.
What does this skill do, and when should you use it?
Launch Monitor covers the T-0 to T+30 product-launch window, checking instrumentation, channel traction, and news echo while the launch is live. It can poll Hacker News rank, points, and comments; Product Hunt votes and featured status; and app-store charts and review metadata. It produces D0, W1, and M1 snapshots, reads spike-versus-sustain and owned capture, and alerts against user-supplied KPI targets. It monitors and informs; it does not make launch-day go/rollback decisions or perform long-term metric diagnosis.
Reads the launch date, tier, stage, and KPI targets; before launch, checks UTM consistency, conversion or signup events, and landing URLs for each launch surface. Uses the referenced Hacker News, Product Hunt, App Store, and GDELT connectors for rank, points, comments, votes, featured status, store charts, review metadata, and news mentions; when a connector is unavailable, asks the user to paste values. Produces polling logs, labels the comments-over-points flamewar signal as an Estimated heuristic, compares D0/W1/M1 actuals with targets, reads spike-versus-sustain and owned-capture, and saves snapshots and handoff summaries under the launch-monitor memory path.
- A product team launches on Hacker News, Product Hunt, and the App Store and needs one launch-window monitoring workflow.
- A growth lead verifies UTM parameters and signup events before launch so channel attribution will be usable.
- A launch owner requests a D0, W1, or M1 snapshot comparing channel actuals with declared KPI targets.
- A community lead wants early warning of unusually comment-heavy Hacker News discussion without delegating launch decisions to the monitor.
- A marketing team closes a T+30 window and prepares peak, sustain, and owned-capture findings for the launch retrospective.
What are this skill's strengths and limitations?
- Covers both pre-launch instrumentation verification and live T-0 to T+30 monitoring.
- Names concrete data paths for Hacker News, Product Hunt, the App Store, and GDELT.
- Degrades to user-pasted values when a connector or key is unavailable.
- Separates Measured, User-provided, and Estimated values and treats the user's own analytics as attribution truth.
- Maintains a clear scope boundary around launch-day decisions.
- Requires a launch date, platforms, and KPI targets; it does not invent target numbers.
- The Product Hunt connector requires a free developer token and has non-commercial API terms, approval, and attribution constraints.
- App-store review text is a manual input rather than an automatically retrieved connector result.
- Platform-reported counts are reference-only; attribution depends on the user's own analytics or store export.
- The supplied source provides no independent test-suite evidence or validation across every possible launch platform.
How do you install this skill?
Install the repository as a collection: in Claude Code run /plugin marketplace add aaron-he-zhu/aaron-marketing-skills, then /plugin install aaron-marketing@aaron; on other Agent Skills-compatible hosts, run npx skills add aaron-he-zhu/aaron-marketing-skills; or clone it with git clone https://github.com/aaron-he-zhu/aaron-marketing-skills. The source does not document a host-specific single-skill installation command for this skill. Its file is launch/prove/launch-monitor/SKILL.md.
How do you use this skill?
Provide a launch date, platforms, and D0/W1/M1 KPI targets. Example: Monitor my launch — we go live 2026-08-01 on HN / Product Hunt / App Store. KPI targets: [D0 / W1 / M1]. For preflight, use Verify my launch instrumentation before 2026-08-01 — here are the launch surfaces and the UTM plan. For a snapshot, use Pull a D0 snapshot: HN rank/points/comments, PH votes, store chart position, news mentions — vs our targets.
How does this skill compare with similar options?
Unlike launch-day-conductor, Launch Monitor observes, polls, and alerts rather than making launch-day action or rollback decisions. Unlike performance-analyzer, rank-tracker, and performance-monitor, it focuses on the product-launch window and launch channels rather than metric deep dives, SEO positions, or monitoring after T+30. Its stated successor at window close is launch-retro-analyzer.