Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/wattdata/plugin/audience-generatenpx skills add wattdata/plugin --skill audience-generategit clone --depth 1 https://github.com/wattdata/pluginWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00182 | $0.05662 |
| Opus 5 | $0.00091 | $0.02831 |
| Sonnet 5 | $0.00036 | $0.01132 |
| Haiku 4.5 | $0.00018 | $0.00566 |
Grade A, and why
audience-generate scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 163 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Generate an audience
Purpose
audience-generate — the build step behind /watt:audience — turns "who I want to reach" into a built audience: the signals behind the brief, scored with the math visible, composed into a set whose reach is measured. The user walks away with a signal stack — the boolean shape of their audience and its measured reach — or a roster: the top groups inside the audience (where those people concentrate), or the set reached by crossing the employment graph to their employers (or to the employees of named companies). Either is ready for the audience-analyze and audience-activate steps.
One spine, many ways to compose. This skill owns the shared spine every build shares — discover the signals, score them, curate the working set, land the audience — and routes by input anchor (what the user has in hand) to the leaf that owns the build:
- a plain-English description of who to reach →
audience-generate-search(today: a size band with greedy, maximum credible reach with broad, precision with lift, the top groups inside the audience with group, or the set reached by crossing the employment graph with traverse — the objective picks). - an owned list — customers, leads, accounts →
audience-generate-list, which resolves the list first (today: resolve-only for a tight matched set, or expand for the widest identity-matched set — both a roster; lookalike, which profiles the list and hands back the signals that define it to build from; or overlay, which scores the resolved list against a pool of signals and ranks it; group and traverse aren't available there).
Route; don't run the compose. Your job at this level is the shared spine below — the language, the build-only lane, and the discover-score-curate-land procedure every leaf composes with. The leaf owns the objective it composes or partitions toward and the strategy procedure that gets there.
Nothing is composed unseen. The user approves the scored working set — and every pivot after it — before any audience is built or measured; the roles, the math, and the target stay on screen the whole way. A finished-looking audience the user didn't steer is this surface's failure mode.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 163 lines · 182 tokens per session scan A ef991fd159f6
audience-generate is a skill published in the GitHub repository wattdata/plugin (5 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 182 tokens to every session and 5,662 once invoked, about $0.0009 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
babysit-pr
Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…
imagegen
Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
next-cache-components-optimizer
Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…