What does this skill do, and when should you use it?
Social Measurement Loop is an organic-social measurement skill for weekly readouts, metric governance, and learning write-back. It locks the denominator used by each engagement-rate definition, rolls up post performance with medians, and keeps organic and boosted activity separate. It can also produce an attributed community-health readout for an owned community and, on request, an executive-only EMV translation labeled as estimated. It does not calculate dollar ROI, determine dark-social methodology, track share of voice, or issue the ECHO gate verdict.
Reads user-exported native analytics from closed platforms, GA4/GSC exports, public keyless connector series, active-channel and cadence data, and prior readouts; builds or loads a period-locked metric dictionary defining ERR, ERI, and ER-by-follower; computes median per-post rollups with organic and boosted content separated; compares the period with prior and baseline results; identifies best and worst performers by channel with one hypothesis each labeled Estimated; optionally computes EMV as an estimated executive translation outside all scores; runs an owned-community readout using Orbit-style distribution, time-to-first-response, and moderator bus factor with employees excluded; produces an explicit keep/stop/try write-back list and handoff summary; and writes the readout to the specified memory location.
- A social lead needs a weekly review of Instagram, Xiaohongshu, or other organic channels with concrete next-cycle actions.
- An analytics team needs one locked denominator dictionary because channels currently report incompatible engagement rates.
- A brand needs to analyze organic posts separately from boosted posts so paid amplification does not distort organic trends.
- A community operator wants Discourse-based health metrics while excluding employee activity from traction and health counts.
- Executives request an EMV translation, but the team must keep that estimate outside scoring and decision trends.
How do you install this skill?
- Confirm that platform export fields, time windows, and denominator-lock records are actually available; otherwise retain NEEDS_INPUT rather than filling gaps from memory or dashboard glances.
- The skill does not provide directly verifiable citations for the external methods behind community-health metrics, EMV, and proxy data; manually review them and preserve the Estimated or proxy labels.
- The skill does not define a complete rollback or cleanup procedure for memory writes; confirm that saved results contain no unnecessary personal or sensitive data.
- Shell / CLI
- Network access
- Local filesystem
Python standard library
To install the repository 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. A plain clone is also supported: git clone https://github.com/aaron-he-zhu/aaron-marketing-skills. The source does not document a standalone installation directory for this skill.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/aaron-he-zhu/aaron-marketing-skills.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/social/observe/social-measurement-loop" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- Run the weekly social readout for the week of 2026-06-29 — here are the Instagram and 小红书 analytics exports plus GA4.
- Build our metric dictionary: which denominator does each engagement rate use per channel, and lock it for future periods.
Provide a period, channels, and exports with a prompt such as: “Run the weekly social readout for the week of 2026-06-29 — here are the Instagram and 小红书 analytics exports plus GA4.” You can also ask: “Build our metric dictionary: which denominator does each engagement rate use per channel, and lock it for future periods.” If required data is missing, the skill should return NEEDS_INPUT with the exact export needed.
What are this skill's strengths and limitations?
- Locks engagement-rate denominators across periods and declares switches as trend restarts.
- Uses median per-post rollups and keeps organic and boosted performance separate.
- Labels figures as Measured, User-provided, Estimated, or proxy.
- Turns best/worst performer findings into an explicit next-cycle write-back.
- Supports employee-excluded community-health analysis for owned communities.
- Closed-platform analysis depends on user exports; the skill does not provide compliant keyless reads for those platforms.
- EMV is an Estimated executive translation and cannot enter scores, trends, or decisions.
- It does not calculate dollar ROI or own dark-social attribution, share-of-voice tracking, or ECHO verdicts.
- The source provides no standalone test-suite or platform-coverage evidence for this individual skill.
- Saving results requires user confirmation, and some memory promotions require permission.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Social Measurement Loop this page | 56 · Use with care | ★ 3k | 12d ago | Apache-2.0 |
| Platform Norm Profiler | 60 · Recommended | ★ 3k | 12d ago | Apache-2.0 |
| Influencer ROI Calculator | 60 · Recommended | ★ 3k | 12d ago | Apache-2.0 |
| Channel Portfolio Planner | 57 · Use with care | ★ 3k | 12d ago | Apache-2.0 |
| Community Participation Warmup Planner | 57 · Use with care | ★ 3k | 12d ago | Apache-2.0 |
Unlike paid-measurement-loop, this skill focuses on organic-social measurement, denominator governance, organic/boosted separation, and social learning write-back. Unlike social-quality-auditor, it produces the measurement readout but does not calculate the ECHO profile result or issue the gate verdict.
How did FollowSkills review this skill?
The skill limits inputs to user-exported analytics and public keyless surfaces, treats imported content as untrusted, prohibits scraping closed platforms, and requires confirmation or authorization for saving, memory promotion, and registry proposals. It does not fully specify sensitive-data minimization, rollback, or storage cleanup within this skill, so 7 points are deducted.
Instructions, scope boundaries, NEEDS_INPUT behavior, source labels, and abnormal-input handling are broadly consistent. Connector availability, key-path reproduction, and skill-specific test evidence are not established; static review therefore cannot exceed 10, with 12 points deducted.
Triggers, Quick Start examples, inputs, outputs, exclusions, and Chinese-language platform examples are clear. Export-format variation, missing-data boundaries, and mainland-China network reachability are not sufficiently documented, and some core connectors depend on the host environment, so 5 points are deducted.
The file provides structured frontmatter, versioning, an Apache-2.0 license, the shared contract, reference material, save-result guidance, and next-skill routing. Maintenance ownership, changelog/update path, installation troubleshooting, and representative output examples remain incomplete, so 5 points are deducted.
The expected deliverable is concrete: a metric dictionary, median rollups, best/worst performers, community-health reporting, and a keep/stop/try write-back. However, no skill-specific output example or execution verification is provided, so direct usability still requires review; the static ceiling is 7 and 1 point is deducted.
Measured/User-provided/Estimated/Calculated labels and references to CHAOSS, Orbit, and formula sources provide some traceability. There is no skill-specific third-party evidence, reproducible example, or test coverage for the key paths; the static ceiling is 5 and 1 point is deducted.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →