HESTIA plugin
HESTIA skills for AI coding assistants. Ask Claude Code, Codex or Antigravity a question about HESTIA data — why a model failed on your Cycle, where an aggregated number came from — and get an answer built from what HESTIA actually computed rather than from guesswork.
This is public. It is published to PyPI as hestia-earth-plugin, and works against the public
HESTIA API, the public data downloads and the public hestia-engine-models repository — no
checkout, no internal access and no credential is required to use it.
pip install hestia-earth-plugin # or: uv tool install hestia-earth-plugin
hestia-plugin install # register the skills with your assistant
Then open your assistant and ask.
Until the first release lands on PyPI, install from the repository instead:
pip install git+https://gitlab.com/hestia-earth/hestia-plugin.git.
The skills
/debug-model |
Why a model failed on a Cycle, Site or ImpactAssessment, or why a value is what it is. Reads the run log the platform stored, checks the model's documented formula against it, and walks the failure upstream to the root cause. |
/debug-aggregation |
Where a number on an aggregation page came from. Rebuilds the weighted mean from the underlying-data download, and names the step — zero-filling, sub-system weighting — that moved it. |
More are being moved across; the Skills page is generated from the skills themselves, so it is always the current list.
Documentation
https://hestia-earth.gitlab.io/hestia-plugin
- Install — every assistant it supports, and where each one reads its skills
- Configure — what a HESTIA API token buys you, and what works without one
- Skills — the full catalogue, generated from the skills themselves
- How it works — the division of labour between the assistant and the tools, and what each skill reads
- Troubleshooting
- Contributing — adding a skill, and what must never be committed
Kept there rather than here so each thing is written down once: the install steps lived in both places until they disagreed.
Working on the plugin
Each skill is a folder in skills/ — a SKILL.md your assistant reads, and any knowledge/
playbooks it consults before diagnosing anything. The deterministic tools they call are in
src/hestia_earth/plugin/, one package per skill; nothing there talks to an LLM.
pip install -e '.[dev,models]'
pytest && black src tests && flake8
hestia-plugin install # re-run after every change to a SKILL.md
To preview a docs change:
scripts/serve-docs.sh # http://127.0.0.1:8000, rebuilding as you edit
scripts/serve-docs.sh --build # one-shot --strict build, the same gate CI runs
It creates a .venv on first run and builds the two generated pages — docs/skills.md and
docs/contributing.md are git-ignored.
Merge requests target master. The rest — commits, playbooks — is in
CONTRIBUTING.md, which is also the source of the Contributing page.
License
MIT — see LICENSE.
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file hestia_earth_plugin-0.1.2.tar.gz.
File metadata
- Download URL: hestia_earth_plugin-0.1.2.tar.gz
- Upload date:
- Size: 99.3 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via:
twine/7.0.0 CPython/3.12.14
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
651d08b5e8bd97dd1968398b9a9221949b6d407033eba2cdc65dc3b2a0b78acd
|
|
| MD5 |
38a59a57f82a15ad9294ab3190feb85b
|
|
| BLAKE2b-256 |
dde591d021e8c7b3458122594442ecb8ddc922036f9ccb14c9db534e575e0afa
|
File details
Details for the file hestia_earth_plugin-0.1.2-py3-none-any.whl.
File metadata
- Download URL: hestia_earth_plugin-0.1.2-py3-none-any.whl
- Upload date:
- Size: 117.9 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via:
twine/7.0.0 CPython/3.12.14
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
b6c700960482a2f9c9662e016898b2b6d91b736339756d95f237f2d6bd479df1
|
|
| MD5 |
b2286cbb90de85303c843e6ddc857283
|
|
| BLAKE2b-256 |
cb1cc0a112248ae530aa8e8e55cca4920fde4216dd458ea3ed4beab2b3f572d5
|