Cover of Fact-checking speech against the source

1920×1080 · 11 s

// Agents and skills //

Fact-checking speech against the source

Data · runs in Claude Code

It takes the word-by-word transcript and whatever sources the project already has: the `pesquisa/fatos.json` fact sheet and/or documents in `.md`, `.txt`, `.csv`, `.json`, and `.pdf`. It outputs `qa/checagem.md` (sorted by severity, for reading) and `qa/checagem.json`, flagging numbers that don't match the source, quotes that don't exist, data that in the source belongs to a different case, misattributions, unsupported claims, and facts the sheet marks as unconfirmed or expired — each item with the timestamp, the sentence as spoken, the source excerpt, and what to do. Any passage that pairs a person's name with an accusatory word gets flagged as sensitive, even when everything else is right. It's a sieve, not a verdict: nothing gets corrected automatically, and the summary lists the documents it couldn't read. For anyone who publishes content with data and quotes and wants to check before, not after.

Data
Runs in Claude Code
Advanced

It takes the word-by-word transcript and whatever sources the project already has: the `pesquisa/fatos.json` fact sheet and/or documents in `.md`, `.txt`, `.csv`, `.json`, and `.pdf`. It outputs `qa/checagem.md` (sorted by severity, for reading) and `qa/checagem.json`, flagging numbers that don't match the source, quotes that don't exist, data that in the source belongs to a different case, misattributions, unsupported claims, and facts the sheet marks as unconfirmed or expired — each item with the timestamp, the sentence as spoken, the source excerpt, and what to do. Any passage that pairs a person's name with an accusatory word gets flagged as sensitive, even when everything else is right. It's a sieve, not a verdict: nothing gets corrected automatically, and the summary lists the documents it couldn't read. For anyone who publishes content with data and quotes and wants to check before, not after.

How to use

01
Copy the prompt
Use the Copy prompt button on this page.
02
Paste it into your AI
Open Claude Code in the project folder and paste the prompt.
03
Install the skill you downloaded
Install the downloaded file in Claude and call the skill in any chat.

In the package

Prompt ready to copy and paste
Skill as a .zip to install in Claude

Goes with

1920×1080 · 11 s

// Agents and skills //

Fact-checking speech against the source

Save

Data · runs in Claude Code

It takes the word-by-word transcript and whatever sources the project already has: the `pesquisa/fatos.json` fact sheet and/or documents in `.md`, `.txt`, `.csv`, `.json`, and `.pdf`. It outputs `qa/checagem.md` (sorted by severity, for reading) and `qa/checagem.json`, flagging numbers that don't match the source, quotes that don't exist, data that in the source belongs to a different case, misattributions, unsupported claims, and facts the sheet marks as unconfirmed or expired — each item with the timestamp, the sentence as spoken, the source excerpt, and what to do. Any passage that pairs a person's name with an accusatory word gets flagged as sensitive, even when everything else is right. It's a sieve, not a verdict: nothing gets corrected automatically, and the summary lists the documents it couldn't read. For anyone who publishes content with data and quotes and wants to check before, not after.

Data
Runs in Claude Code
Advanced

Available on the Pro plan

Try now for 99 cents

See sample prompt

Prompt

Cole no Claude Code, na pasta do projeto

Pro

Use this template to build your piece with Mosaicraft.

Start with the format and the length of the style.
Then describe the scene, line by line:
"the first line, the one that holds attention"
"the second line, with the highlighted bit"
Keep the colors, the type, and the rhythm of the style.
At the end, show the result full screen.
Match the tone to the people who will watch.

Deliver a file that is ready to use.

This skill is on the Pro plan

Assine pra copiar este e todos os outros prompts. Item novo toda semana.

See the free ones

Then $8.88 every 28 days.

Already subscribed? Log in

Try now for 99 cents

It takes the word-by-word transcript and whatever sources the project already has: the `pesquisa/fatos.json` fact sheet and/or documents in `.md`, `.txt`, `.csv`, `.json`, and `.pdf`. It outputs `qa/checagem.md` (sorted by severity, for reading) and `qa/checagem.json`, flagging numbers that don't match the source, quotes that don't exist, data that in the source belongs to a different case, misattributions, unsupported claims, and facts the sheet marks as unconfirmed or expired — each item with the timestamp, the sentence as spoken, the source excerpt, and what to do. Any passage that pairs a person's name with an accusatory word gets flagged as sensitive, even when everything else is right. It's a sieve, not a verdict: nothing gets corrected automatically, and the summary lists the documents it couldn't read. For anyone who publishes content with data and quotes and wants to check before, not after.

How to use

01
Copy the prompt
Use the Copy prompt button on this page.
02
Paste it into your AI
Open Claude Code in the project folder and paste the prompt.
03
Install the skill you downloaded
Install the downloaded file in Claude and call the skill in any chat.

In the package

Prompt ready to copy and paste
Skill as a .zip to install in Claude

Goes with

ProLooking at the camera or readingMarks, segment by segment, where the person looks at the camera, where they're reading, and where there's no face in the frame, with timestamps.PythonProductionIntermediate

ProScript in your speaking styleMeasures how you talk from your past transcripts and writes the script in that style, in two files: the clean narration and the editing script.ScriptResearchIntermediate

ProWord-by-word transcriptionTranscribes audio and video (Portuguese by default) with the timing of every word, installs Whisper on its own, and handles hour-long videos in chunks.PythonCaptionsIntermediate

Available on the Pro plan

Try now for 99 cents

See sample prompt

Prompt

Cole no Claude Code, na pasta do projeto

Pro

Use this template to build your piece with Mosaicraft.

Start with the format and the length of the style.
Then describe the scene, line by line:
"the first line, the one that holds attention"
"the second line, with the highlighted bit"
Keep the colors, the type, and the rhythm of the style.
At the end, show the result full screen.
Match the tone to the people who will watch.

Deliver a file that is ready to use.

This skill is on the Pro plan

Assine pra copiar este e todos os outros prompts. Item novo toda semana.

See the free ones

Then $8.88 every 28 days.

Already subscribed? Log in

Try now for 99 cents

How to use

Copy the prompt

Paste it into your AI

Install the skill you downloaded

In the package

Goes with

This skill is on the Pro plan

How to use

Copy the prompt

Paste it into your AI

Install the skill you downloaded

In the package

Goes with

This skill is on the Pro plan