Free skills
Proposed Three skill drafts, each written for one chore over the pauhu.ai connector. The results below are what each draft aims for, and none is proved on the connector yet. Each one shows its plan first and changes nothing until you approve. No sign-up is needed to read or copy them. Connector use follows its own limits, shown on your first request.
Connector: https://api.pauhu.ai/mcp. Skill licence: not set yet
Check that citations are current
Tax and compliance teams ask whether a source is current and whether a later change replaced it.
Input: You give a draft that cites laws or guidance.
Intended result: A table with one row per citation: what the draft cites, whether a later text changes it, and a link to the original. If nothing is found, the row says so.
---
name: check-citations
description: Check that citations are current. Proposed draft. Shows a plan and changes nothing until approved.
---
# Check that citations are current
Status: proposed. This result is not proved on the connector yet.
Input: You give a draft that cites laws or guidance.
Intended result: A table with one row per citation: what the draft cites, whether a later text changes it, and a link to the original. If nothing is found, the row says so.
Connector: https://api.pauhu.ai/mcp
Steps
1. List every citation found in the draft and show the list.
2. Wait for approval of the list.
3. Ask the connector one question per approved citation, with the date the draft relies on.
4. Return the table. Leave the draft unchanged.
5. Edit the draft only if the reader approves each edit.
Find conflicting terms
Localization teams ask how to deal with stale terms, conflicting terms and duplicate records.
Input: You give a glossary or term list.
Intended result: A list of terms that conflict or repeat, with a sourced suggestion for each. If no source is found, the entry says so.
---
name: check-terms
description: Find conflicting terms. Proposed draft. Shows a plan and changes nothing until approved.
---
# Find conflicting terms
Status: proposed. This result is not proved on the connector yet.
Input: You give a glossary or term list.
Intended result: A list of terms that conflict or repeat, with a sourced suggestion for each. If no source is found, the entry says so.
Connector: https://api.pauhu.ai/mcp
Steps
1. List the duplicates and conflicts inside the file and show them.
2. Wait for approval of which terms to look up.
3. Ask the connector one question per approved term.
4. Return the list. Leave the file unchanged.
5. Edit the file only if the reader approves each edit.
Summarize a dataset's terms
Teams buying data ask what use is allowed, what happens after the contract ends and how often the data is refreshed.
Input: You give the published terms of a dataset you are considering.
Intended result: A short summary of permitted uses, rights after termination and refresh schedule, each point linked to the passage. A point the terms do not cover reads "not stated". It is a reading aid, not legal advice.
---
name: read-dataset-terms
description: Summarize a dataset's terms. Proposed draft. Shows a plan and changes nothing until approved.
---
# Summarize a dataset's terms
Status: proposed. This result is not proved on the connector yet.
Input: You give the published terms of a dataset you are considering.
Intended result: A short summary of permitted uses, rights after termination and refresh schedule, each point linked to the passage. A point the terms do not cover reads "not stated". It is a reading aid, not legal advice.
Connector: https://api.pauhu.ai/mcp
Steps
1. List the questions it will answer and show them.
2. Wait for approval of the list.
3. Read the terms you gave and ask the connector only about terms the text cites.
4. Return the summary. Change nothing.