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 skills add swan-gtm/gtm-skills --skill ai-visibility-prompt-researchgit clone --depth 1 https://github.com/swan-gtm/gtm-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/swan-gtm/gtm-skills/ai-visibility-prompt-research)<a href="https://agentmods.dev/skills/swan-gtm/gtm-skills/ai-visibility-prompt-research"><img src="https://agentmods.dev/badge/skills/swan-gtm/gtm-skills/ai-visibility-prompt-research/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/swan-gtm/gtm-skills/ai-visibility-prompt-research"><img src="https://agentmods.dev/badge/skills/swan-gtm/gtm-skills/ai-visibility-prompt-research.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What 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 5.1 | $0.00184 | $0.01015 |
| Opus 5 | $0.00092 | $0.00508 |
| Sonnet 5 | $0.00037 | $0.00203 |
| Haiku 4.5 | $0.00018 | $0.00102 |
Grade A, and why
ai-visibility-prompt-research 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 9d ago.
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 — 46 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AI visibility prompt research
A monitoring prompt set decides what you can even see. Anchor it on the brand's own marketing copy and you measure prompts nobody searches; anchor it on real market demand and you measure whether the brand wins the questions buyers actually ask an assistant. The whole discipline is holding the brand back until the end so the research stays grounded in the market.
The play
- Frame — no brand yet. Collect the category (plain description, no brand name), the named competitor set, the target regions and their languages, and which assistants you'll monitor. Say up front that you're withholding the brand on purpose.
- Harvest category language, brand last. Pull real buyer phrasing from three wells: keyword and category terminology (ask for any existing keyword or targeting file, then expand from the open web); real questions and pain points mined from communities, forums, Q&A, and "alternatives / best-of" comparison content; and — asked after the open-web pass — first-party sales-call and pre-sales transcripts, the highest-signal source. Tag every item with honest provenance. (See
references/open-web-sourcing.md.) - Synthesize personas and the non-branded prompt set. Merge all signals into 4–7 personas with their pain points and the solutions they seek, in category terms. Map each to an awareness stage, weighting consideration and decision over pure curiosity. Cluster into a handful of topics. Write each prompt the way a real buyer asks an assistant — a genuine question, not a keyword string — in the target region's language (translate, don't just localize).
- Brand pass — last, and isolated. Only now introduce the brand. Add competitive comparisons, "alternatives to", pricing, and fit questions as their own dedicated branded topic, flagged branded, kept separate from the category topics.
- Assemble, validate, hand off. Build one file to your platform's importer contract, validate it (valid enums, one language per row matching its regions, no duplicates, sane per-topic and per-engine spread), show the distributions, and hand it to the user to upload. Never auto-import. (See
references/prompt-file-contract.md.)
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 9d ago First seen · 46 lines · 184 tokens per session scan A a94378da3d0e
ai-visibility-prompt-research is a skill published in the GitHub repository swan-gtm/gtm-skills (150 stars, last pushed 2d ago), licensed MIT. It adds 184 tokens to every session and 1,015 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 2026-09-03.
Other skills, from other repositories
c2pa-metadata
Embed a C2PA provenance manifest into an AI-generated marketing asset (PNG, JPG, WebP, GIF, TIFF, MP4, MOV, WebM, MP3, WAV, PDF) via scripts/embed-c2pa.py — produces a signed copy of the file carrying IPTC digital-source-type AI claims, an optional c2pa.ai-disclosure assertion for EU AI Act Article 50 (applicable 2…
multilingual-score
Score translated or localized content into one composite quality verdict — technical translation scoring (length ratio, formatting, do-not-translate terms, placeholders), content quality in the target language, brand voice consistency, and target-market compliance — classified as Publish Ready (85+), Native Speaker…
attribution-model
Design a multi-touch attribution strategy — recommends the best-fit model for the business's sales cycle and data maturity, defines credit-distribution rules and lookback windows, maps platform-specific setup (GA4, HubSpot, Salesforce, warehouse), and documents tracking gaps and known blind spots. Triggers on…
rank-monitor
Set up and run keyword ranking monitoring — baseline capture, scheduled position checks against GSC and connected rank-tracker MCPs, and severity-tiered alerts (minor/major/critical) on drops; --features adds a query-by-SERP-feature ownership matrix including AI Overview citation presence. Triggers on…
add-integration
Walk through adding a custom MCP server integration to the plugin — searches npm for an existing MCP package (or scaffolds a custom server from the plugin's guide), generates the exact .mcp.json entry, sets up environment-variable credentials, tests connectivity, and documents the tools the new server exposes.…
case-study-plan
Build a complete case-study creation blueprint — a Challenge-Solution-Results narrative framework, 15-20 client interview questions plus 10 internal-team questions, a data-visualization plan, format specifications (PDF, web page, slide deck, video script outline, social snippets, sales one-pager), a distribution…