Loading skill
Loading skill
moatlight-read-reportMoatlight · published in Moatlight/moatlight-agent-kit
Turns a finished report into a short, honest summary and the next three changes, per engine and in your buyers' words.
Agents keep only this description in mind and load the skill when a request matches it. It is the exact text from SKILL.md.
Read a finished Moatlight report and turn it into a short, honest summary and a ranked to-do list: how often AI named the business, which brands it named instead, and the next three changes. Use when a report has finished, or when the user asks what their AI visibility results mean, who beats them in ChatGPT answers, or what to change first.
Stops if the report has not finished. A partial report is never summarized.
Named in n of m answers for ChatGPT and for Google AI Overview, never one averaged number.
Each with why, what to do and where. Blanks in a draft stay blank until you fill them.
After installing, say it in your own words. For example:
Everything the skill contains. Each file opens as plain text on this site; the same files are on GitHub.
---
name: moatlight-read-report
description: "Read a finished Moatlight report and turn it into a short, honest summary and a ranked to-do list: how often AI named the business, which brands it named instead, and the next three changes. Use when a report has finished, or when the user asks what their AI visibility results mean, who beats them in ChatGPT answers, or what to change first."
license: MIT
compatibility: "Needs network access to moatlight.com. Works in any agent that can make HTTPS calls or use an MCP server."
metadata:
author: Moatlight
version: "1.2.1"
homepage: https://moatlight.com/docs
---
# Read a Moatlight report
## Get the data
- `GET https://moatlight.com/api/v1/reports/{id}` (scope `reports:read`) → `stage` and, when `finished`, `summary`: `answers`, `named`, `share_of_voice`, `rank`, `per_engine`, `top_brands`, `lost_questions` and, only when present, `sentiment` (how the answers that named the business described it).
- `GET https://moatlight.com/api/v1/reports/{id}/actions` → `moves` (up to three) and `lost_questions`.
- Do not know the id? `GET https://moatlight.com/api/v1/reports?limit=10`.
If `stage` is not `finished`, say so and stop. Do not summarize a partial report.
## Answer in this shape
```
<Business> was named in <named> of <answers> answers (ChatGPT <n>/<m>, Google AI Overview <n>/<m>).
Most named instead: <top brand> in <k> answers.
Questions it lost: <2 or 3 from lost_questions, in the user's words>
Tone (only if sentiment is present): <positive.k> of <positive.n> graded answers described it positively. Say it is an automated reading, and if test_grader is true say the labels are a test.
Next three changes
1. <title> (effort: <effort>)
Why: <why>
Do: <todo>, in <where>
2. ...
```
## Gotchas
- Moves may include a `draft` with [blanks]. Never fill a blank with a fact you do not have. Ask the user.
- `share_of_voice` and `rank` are relative to the brands seen in this report only.
- Engines differ. Report each one. Do not average them into one number.
- `moves` can be empty when the business was named in nearly every answer. Say that, and do not invent changes.
- A re-run after 30 days shows what moved. `recheck` repeats the report.
## Do not
- Do not say a change "will" raise visibility. The report states a hypothesis.
- Do not present counts as traffic or revenue.
Found a problem in this skill? Open an issue on GitHub or write to Moatlight.
