Content
82%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-structured agent skill body: lean, highly actionable CLI and math guidance, clearly sequenced workflows with embedded checkpoints, and well-signaled one-level-deep references. The main weaknesses are a redundant Voice section, missing fallback guidance when rescue-mode data is incomplete, and bundle references that don't resolve in the provided file tree.
Suggestions
Verify or ship the referenced bundle (scripts/webinar_funnel_scorer.py, references/*.md, templates/webinar-plan-template.md) alongside this SKILL.md — as provided, every reference and CLI command points at files that do not exist in the bundle.
Add a fallback step to Mode 2 for when actual funnel numbers are missing or partial (e.g., estimate from list size and benchmarks, then flag the diagnosis as provisional).
Collapse the four Voice quotes into the two most distinctive behaviors — the persona is already summarized in the description and duplicated in Success Metrics.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and assumes Claude's competence — no explaining what a webinar or funnel is, concrete worked funnel math, tight CLI docs. However, the four-quote Voice section (~15 lines) largely duplicates the voice clause already in the description, and the 'Treats a webinar as a funnel, not an event' framing is repeated across sections. Not 5: minor redundancy that could be trimmed; not 3: there is no over-explanation of concepts Claude already knows. | 4 / 5 |
Actionability | Copy-paste-ready bash invocations for the funnel scorer (file, stdin, and demo modes), a complete input JSON example with the audience enum and required/optional fields, a worked backward-sizing example with realistic rates, and a named template for deliverables. Not 4: the CLI and output documentation cover the common cases completely. | 5 / 5 |
Workflow Clarity | Three clearly sequenced modes; Mode 1 embeds an explicit reality-check checkpoint (step 3: 'if required visits exceed reachable audience, fix goal/format/budget now') and Mode 2 is a score→diagnose→fix loop anchored by the funnel scorer. Not 5: Mode 2 offers no guidance when actual funnel numbers are missing or partial, and Mode 3 (evergreen) has no validation checkpoint at all; not 3: checkpoints in Modes 1–2 are explicit, not implicit. | 4 / 5 |
Progressive Disclosure | Good structure: the Skill Integration section enumerates every bundle file (script, three references, template) with one-line descriptions, references are one level deep, and they are linked at point of need within the workflows. Not 5: none of the referenced paths (marketing-skill/skills/webinar-marketing/...) exist in the bundle under review, so the navigation points at files that cannot be verified; not 3: the references are clearly signaled and well organized, not buried. | 4 / 5 |
Total | 17 / 20 Passed |