kedro-skills
Distribute AI coding skills to Kedro projects. One command installs contextual guidance that activates automatically when your AI assistant edits matching files — across Cursor, GitHub Copilot, Claude Code, and Codex CLI.
Installation
pip install kedro-skills
Quick start
# Inside any Kedro project:
kedro skills list
kedro skills install catalog-config
This writes managed files into your project. They should be committed to git so the whole team benefits.
What it writes
my-kedro-project/
├── .agents/skills/catalog-config/SKILL.md # canonical (Agent Skills standard)
├── AGENTS.md # content discovery (Cursor, Copilot, Codex CLI)
├── .cursor/rules/catalog-config.mdc # glob-scoped activation (Cursor)
├── .github/instructions/catalog-config.instructions.md # glob-scoped activation (Copilot)
└── .claude/skills/catalog-config/SKILL.md # Claude Code (inline copy, always discoverable)
AGENTS.md is the content-discovery channel — one file reaches Cursor, Copilot,
and Codex CLI. .cursor/rules/ and
.github/instructions/ add glob-scoped activation so the skill fires only when
editing matching files, not on every prompt.
CLI reference
| Command | Description |
|---|---|
kedro skills list |
Show available skills and their install status |
kedro skills install <id> |
Install a skill for all supported IDEs |
kedro skills install --all |
Install every available skill |
kedro skills install --ide cursor,claude |
Restrict to specific IDEs |
kedro skills install --force |
Overwrite even if files were hand-edited |
kedro skills update |
Re-render all installed skills (picks up package upgrades) |
kedro skills update --force |
Overwrite modified files during update |
kedro skills uninstall <id> |
Remove a skill and all its managed files |
kedro skills uninstall --force |
Remove even if files were hand-edited |
IDE-specific notes
- Claude Code: Reads
.claude/skills/<id>/SKILL.mddirectly. The skill is always discoverable (nopaths:scoping in the Claude copy). If you wantAGENTS.mdcontent in Claude sessions too, add@AGENTS.mdto aCLAUDE.mdfile. - Cursor:
.cursor/rules/*.mdcfires only when editing files matchingglobs:patterns. - GitHub Copilot:
.github/instructions/*.instructions.mdfires only when editing files matchingapplyTo:patterns. - Codex CLI: Reads
AGENTS.mdnatively. Skill block is always active.
How to try it out
# Install kedro-skills from source (in the kedro-skills repo root)
pip install .
# Create a test Kedro project
pip install kedro
kedro new --name test-project -s spaceflights-pandas
cd test-project
# See what's available
kedro skills list
# Install the catalog-config skill
kedro skills install catalog-config
# Check what was written
ls .agents/skills/catalog-config/SKILL.md
cat AGENTS.md
ls .cursor/rules/
ls .github/instructions/
ls .claude/skills/catalog-config/
# Running install again is safe — nothing changes
kedro skills install catalog-config --ide cursor,copilot,claude
# If you hand-edit a managed file, update will let you know
echo "modified" > .cursor/rules/catalog-config.mdc
kedro skills update # refuses and names the modified file
kedro skills update --force # overwrites your edit with the managed version
# Clean removal when you no longer need a skill
kedro skills uninstall catalog-config
kedro skills list # back to "not installed"
Telemetry
When kedro-telemetry
is installed (it ships with Kedro), the install, update and uninstall
commands report anonymous usage events:
| Event | Properties |
|---|---|
kedro_skills_install |
skill_id, target_ides (sorted, comma-separated), install_all (whether --all was used), success |
kedro_skills_update |
skills_updated (count), drift_detected (whether locally modified skill files were found), success |
kedro_skills_uninstall |
skill_id |
Skill ids are reported when they match a skill shipped with this package;
anything else is sent as unknown. kedro skills list sends no event. No
skill content, file paths, project details or personal data are ever sent.
The standard
Kedro telemetry opt-out
(DO_NOT_TRACK, KEDRO_DISABLE_TELEMETRY, or a .telemetry file with
consent: false) disables these events too.
Development
See README_DEV.md for setup, running tests, authoring skills, and releasing.
Links
Metadata
Release files for kedro-skills 0.2.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| kedro_skills-0.2.0.tar.gz | 23.2 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| kedro_skills-0.2.0-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 52.0 kB
Release files / kedro_skills-0.2.0.tar.gz
| Download URL | kedro_skills-0.2.0.tar.gz |
|---|---|
| Size | 23.2 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
01f006d8462bf621c714b978dae2d21fd7bd7d43ef78decec0764345e088253e
|
|
BLAKE2b-256 checksum How to use checksums |
069c9bd91b7538590951299d9cae7ace84d343af43ad160e36bdc180f94c6d33
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 3, 2026.
Transparency logRelease files / kedro_skills-0.2.0-py3-none-any.whl
| Download URL | kedro_skills-0.2.0-py3-none-any.whl |
|---|---|
| Size | 28.8 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
76853ad4bf271f5d51eaa7faafc398c18d8f78f612fee5ca5c4bfb0afd2a8402
|
|
BLAKE2b-256 checksum How to use checksums |
4e12819d1572843f9bf71c10cd59e90886c6c4da04fa28faa774f2bd774ad596
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 3, 2026.
Transparency log