evaluate-content

Installation
SKILL.md

Content Eval — Orchestrator

Evaluation skill. Converts published organic-content evidence into a cycle snapshot + ledger row + narrowly-scoped next action inside an existing eval loop. One primary platform per cycle; secondary platforms are context only.

Core Question: "Did this content cycle, on its primary platform, create measurable signal strong enough to keep / discard / watch / block — and what should the next strategy/execution skill know?"

Why, methodology, history: references/playbook.md [PLAYBOOK]. Capability metadata (route triggers, prerequisites, load map): routing.yaml.

Critical Gates

  1. Existing eval loop required. program.md + context.md absent → NEEDS_CONTEXT, recommend /run-pipeline. This skill does not create loops.
  2. Organic non-video content only. Text / image / carousel posts. Short-form video → evaluate-shortform; paid-ad → evaluate-ad. Otherwise → NEEDS_CONTEXT, route to sibling.
  3. Measurement evidence + primary metric decides the row. Need ≥1 metric source, window, and current value for the loop's primary metric (engagement rate · save rate · CTR · conversion rate — operator-pick via program.md). Secondary metrics explain diagnosis; they don't override unless program.md defines a guardrail failure.
  4. One primary platform per cycle. Secondary platforms live in Cross-Platform Context subsection — they inform diagnosis but DO NOT drive the verdict. A 9-platform campaign = 9 separate cycles, one primary platform each.
  5. No fabricated analytics. Unknown values stay unknown. Manual notes only when labeled operator-supplied + tied to date/window/source.
  6. Attribution confidence must be explicit. Every verdict includes sample size (impressions/reach + window), baseline comparability (same platform, same content type, comparable window), confounders (algorithm change, posting-time shift, follower-count change, cross-post cannibalization), and confidence: high | medium | low | blocked.
  7. Evaluation does not generate content. Recommend changes; route copy authorship to write-social, distribution to publish-social, visual assets to produce-asset.
Installs
24
GitHub Stars
12
First Seen
May 24, 2026
evaluate-content — hungv47/meta-skills