audience-analyze

Read who a built audience reaches, as aggregates over a deterministic sample — and, on request, a shareable signal membership report. The read step behind /watt:audience; routes to the way in that fits what the user has — a brief, signals they already hold, or a list of people. Aggregates only — never individual records, never contact data, never an ad-platform export. Not a user command — /watt:audience is the front door. Use when a read-shaped ask arrives — "who's actually in this audience", "what do these people look like", "an audience profile for my client" — or to sanity-check a signal stack before exporting.

Quality

84%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

73%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A well-structured, actionable canon with a clearly sequenced procedure and explicit failure handling. The main weakness is restated routing across three sections that could be tightened.

Suggestions

Consolidate the routing description: "Purpose", "Works with", and "Entry — pick the way in" each restate the same three ways in — keep the canonical list once and reference it from the other sections.

Inline a one-line summary of mode A vs mode B profiling inputs so a leaf can handle the common case without round-tripping to context/profiling.md, reserving the bundle for full detail.

Dimension	Reasoning	Score
Conciseness	Dense and purposeful with no padding of concepts Claude already knows, but routing is restated three times — "Purpose", "Works with", and "Entry — pick the way in" each list the same three ways in — which could be consolidated.	3 / 5
Actionability	Provides a copy-paste-ready bash invocation for build_report_membership.py with full flags plus concrete cat commands for bundle files; the core profiling run is delegated to the bundle rather than inlined, leaving a minor gap.	4 / 5
Workflow Clarity	The read & report procedure is numbered (1 run inline, 2 render, 3 end at decision) with explicit error-recovery checkpoints — "never fill a dashboard band with plausible numbers", the auth liveness gate, and routing to generate on an empty set.	5 / 5
Progressive Disclosure	Well-organized into labeled sections and defers detail to one-level-deep bundle references (context/profiling.md, context/record.md) and a real script (build_report_membership.py, present in scripts/); the inlined shared canon is large but appropriate for leaves to compose with.	4 / 5
	Total	16 / 20 Passed

Description

87%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, concrete description with explicit what/when trigger guidance and a clearly distinguished aggregates-only lane. It is slightly domain-jargon-laden but comprehensive and actionable.

Dimension	Reasoning	Score
Specificity	Names concrete actions — "Read who a built audience reaches, as aggregates", "a shareable signal membership report", "routes to the way in" — listing several specific actions, but the domain vocabulary ("signal membership report") keeps it just short of comprehensive coverage.	4 / 5
Completeness	Explicitly states what ("Read who a built audience reaches, as aggregates over a deterministic sample — and ... a shareable signal membership report") and when ("Use when a read-shaped ask arrives" with concrete quoted trigger phrases), satisfying the top anchor on both axes.	5 / 5
Trigger Term Quality	Includes natural user phrasings as triggers — "who's actually in this audience", "what do these people look like", "an audience profile for my client" — good coverage, though the terms are domain-specific rather than broadly synonym-rich.	4 / 5
Distinctiveness Conflict Risk	Carves a clear niche — "Aggregates only — never individual records, never contact data, never an ad-platform export" and "Not a user command — /watt:audience is the front door" — explicitly distinguishing it from generate and activate with minimal conflict risk.	5 / 5
	Total	18 / 20 Passed

Validation

100%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: wattdata/plugin
Path: skills/audience-analyze/SKILL.md
Commit: 880e260

Reviewed: 4 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.