CtrlK
BlogDocsLog inGet started
Tessl Logo

pyxll/pyxll-agent-skills

A curated collection of Agent Skills for working with PYXLL, to help AI agents write and understand code using the PyXLL Excel add-in.

99

1.56x
Quality

90%

Does it follow best practices?

Impact

100%

1.56x

Average score across 17 eval scenarios

SecuritybySnyk

Advisory

Suggest reviewing before use

Overview
Quality
Evals
Security
Files

SKILL.mdskills/fetch-pyxll-docs/

name:
fetch-pyxll-docs
description:
Fetch the PyXLL documentation and use it as context for any task involving PyXLL. Invoke automatically before writing, editing, reviewing, or debugging any PyXLL code, and for release notes, changelog, or version history lookups. Training-data knowledge of PyXLL is incomplete and may be wrong — the live docs are the only authoritative source.
user-invocable:
No
metadata:
{"author":"pyxll","version":"0.1.4"}

Fetch the PyXLL documentation and use it as context for any PyXLL task. Training-data knowledge of PyXLL is incomplete and may be wrong — the live docs are the only authoritative source.

Steps

  1. Fetch the documentation index to find relevant pages:

    curl -s https://www.pyxll.com/llms.txt

    Read the entire output without truncating — the index contains a navigation guide mapping common tasks to sections, followed by page titles, descriptions, and URLs grouped by topic.

  2. Fetch the individual pages relevant to the task directly by their URL:

    curl -s <page-url>
  3. If the index does not surface what you need, use the search script to find pages by keyword.

    The search script is located at scripts/search-pyxll-docs.sh relative to this skill's base directory.

    <SKILL_BASE_DIR>/scripts/search-pyxll-docs.sh <keyword> [keyword2 ...]

    Important: Replace <SKILL_BASE_DIR> with the actual base directory path provided when this skill is loaded.

    The script outputs a list of matching page URLs (one per line). Fetch each returned URL individually using curl.

  4. If the task involves writing or modifying code that uses the Excel COM API — any code that calls pyxll.xl_app() or XLCell.to_range(), or that accesses COM objects such as Range, Worksheet, Workbook, Font, Interior, Borders, etc.:

    • Invoke the pywin32-excel-docs skill before writing any code. Its resources/ directory is the authoritative reference for all Excel COM classes — do not rely on training data for COM API details.
    • Fetch https://www.pyxll.com/docs/userguide/vba.md and read it in full — it documents critical differences between VBA and Python, including how COM properties that take arguments must be called as Get<PropertyName>(args) rather than Property(args) as in VBA.
  5. If the task involves writing or modifying a ribbon or context menu XML file (customUI XML, ribbon = in pyxll.cfg, or any onAction/getLabel/getEnabled ribbon callback):

    • Invoke the office-customui-xsd skill before writing any XML. It is the authoritative reference for which elements, attributes, and nesting are valid — training-data knowledge of the schema is incomplete and produces invalid XML.

Rules

  • ALWAYS fetch these docs before writing, modifying, or troubleshooting any PyXLL-specific code or behaviour. Before suggesting a manual workaround for a PyXLL problem, check whether PyXLL already has a built-in solution (decorator parameter, config key, or feature).
  • Before using any PyXLL class, function, decorator, or configuration setting (including pyxll.cfg section names and their keys), fetch the relevant documentation and use only what is explicitly documented. Never infer behaviour, key names, or parameter names from conventions or assumptions — if it is not in the docs, do not use it.

README.md

tile.json