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 Citlyze/citlyze-skills --skill citlyze-prompt-auditgit clone --depth 1 https://github.com/Citlyze/citlyze-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/citlyze/citlyze-skills/citlyze-prompt-audit)<a href="https://agentmods.dev/skills/citlyze/citlyze-skills/citlyze-prompt-audit"><img src="https://agentmods.dev/badge/skills/citlyze/citlyze-skills/citlyze-prompt-audit/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/citlyze/citlyze-skills/citlyze-prompt-audit"><img src="https://agentmods.dev/badge/skills/citlyze/citlyze-skills/citlyze-prompt-audit.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00063 | $0.00633 |
| Opus 5 | $0.00032 | $0.00316 |
| Sonnet 5 | $0.00013 | $0.00127 |
| Haiku 4.5 | $0.00006 | $0.00063 |
Grade A, and why
citlyze-prompt-audit 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 11d 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 — 37 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Citlyze prompt audit
Review the prompt portfolio: find prompts where the brand is invisible, engines or locations that drag performance down, and intent gaps worth tracking next.
Prerequisites
The Citlyze MCP server must be connected (setup guide). Tools are read-only; adding or editing prompts happens in the app, not over MCP.
Workflow
- Inventory. Call
list_promptswithactive: true. Each prompt hasquery_id,query_text,intent,tier,topic, andprompt_group. Summarize the portfolio: counts by intent and topic. - Baseline. Call
get_visibility_overviewfor the latest completed window to know the brand's average per-engine performance. - Drill into suspect prompts. For prompts the user cares about (or a sample of high-
tierones), callget_prompt_visibilitywith thequery_id. Rows inby_engine_locationbreak metrics down per engine and location; look for:- prompts with near-zero
mention_rateeverywhere (invisible) - prompts strong on one engine but absent on another (engine gap)
- prompts that diverge sharply by location (localization gap)
Read every conclusion alongside
n_runs; thin run counts make weak evidence.
- prompts with near-zero
- Coverage gaps. Compare the prompt inventory against how buyers actually ask: are commercial-intent prompts ("best X for Y", "X vs Y", "X alternatives") represented for each core topic? Which competitor comparisons are missing? Which customer questions from the user's domain knowledge have no tracked prompt?
- Deliver:
- Weak prompt table: prompt text, worst engine/location, mention rate, n_runs, suspected reason.
- Suggested prompts to add (5-15), each with intent label and the reason it earns a slot. Remind the user to add them in the Citlyze app.
Interpretation rules
- Treat
intent,tier,topic, andprompt_groupas the workspace's own labels; group by whatever values are present rather than assuming a fixed set. - Zero visibility on a prompt is not automatically bad tracking; it may be the honest baseline the user needs to see.
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.
- 11d ago First seen · 37 lines · 63 tokens per session scan A 4c181b7e6432
citlyze-prompt-audit is a skill published in the GitHub repository Citlyze/citlyze-skills (0 stars, last pushed 1mo ago), licensed MIT. It adds 63 tokens to every session and 633 once invoked, about $0.0003 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
ganhuo-geo-engineer
Use this Skill when the user provides an existing article, product page, tutorial, FAQ, or knowledge note and wants to rebuild it into a GEO or AI-search-friendly content asset. Use it for old-content refresh, citation-readiness improvement, answer-first restructuring, GEO upgrades, and Ganhuo AI content workflows. Do…
xerj-code
Reference-coding with XERJ. Clone the libraries that already solved your problem, index them locally, and retrieve the exact implementation before writing code — so the agent reads passages instead of re-deriving algorithms across retry loops. Use when starting a task in an unfamiliar API, porting an algorithm, or…
fire-your-seo-agency
A procedure for improving how a website appears in search engines and how AI answer systems find and cite it. It covers search, answer-engine, generative-AI, and Naver visibility.
geo-loop
Run one bounded eGEOagents loop iteration over a workspace domain - read the charter and fresh collector data, do ONE unit of work, write substrate artifacts, append one Timeline entry and one LOG line. Use for loop mode, /geo:loop, scheduled GEO runs, or continuous monitoring.
competitive-analysis
Analyze AI-search competitors for a query and recommend ranking strategy. Use when users ask competitor analysis, who ranks, or competitive landscape.
validation-doctor
Check Brave Search and Chrome DevTools MCP availability and provide exact setup snippets. Use when validation dependencies are missing or uncertain.