Conda Project Environment Manager

Conda Project Environment Manager

Use this skill when the repo needs one stable helper for managing Python/conda environments across scripts, research utilities, and infrastructure workflows.

When to use

Use this skill when:

1. the project needs to know which conda installations and envs are available,
2. a workflow needs a reproducible env choice for scripts such as model fitting, PDF processing, or slide generation,
3. packages are missing from the default shell Python and the project should standardize conda run -n ... instead of ad hoc shell state,
4. multiple conda roots (miniconda, anaconda) coexist and env provenance is becoming confusing,
5. env drift, missing packages, or conflicting Python roots are blocking project work.

Scope boundary

This skill owns:

• project-level conda env discovery and reporting,
• explicit env-selection guidance for scripts and workflows,
• package-availability checks for target envs,
• environment snapshot artifacts for later maintenance or debugging,
• minimal reproducibility guidance for conda run -n ... use in this repo.

It does not own:

• workflow routing or workflow ownership,
• domain experiment conclusions,
• arbitrary package installation policy for unrelated user machines,
• replacing system package managers outside project needs.

Preferred usage pattern

1. discover conda roots and envs,
2. record package availability in a snapshot artifact,
3. choose one explicit env per workflow/tool family,
4. call scripts through conda run -n <env> ... when reproducibility matters,
5. update the snapshot when environments change materially.

Expected outputs

Typical outputs include:

1. *.envs.tsv
2. *.summary.md
3. a short recommendation on which env should back which script family
4. maintenance-log or KB refresh only when the environment contract changes materially

Hard rules

1. Do not rely on accidental shell activation as the default reproducibility mechanism.
2. Prefer explicit conda run -n <env> when a script depends on non-system packages.
3. Keep the number of project-preferred envs small; do not create a new env for every one-off task.
4. Record when both miniconda and anaconda roots coexist so future users do not guess which one is active.

Operational references

1. Load references/background-knowledge-links.md for the current environment topology and project conventions.
2. Load references/update-mechanism.md when environment structure or preferred env ownership changes.
3. Use scripts/common/conda_project_env_report.py to generate the canonical snapshot before recommending a new env binding.