ci

Custom agent imported from eco-emma/emma_model (.github/agents/ci.agent.md). Copyright stays with the author.

Orchestrate a full pre-commit review of the codebase. Run each sub-check in order, collect results, and produce a single summary report.

Steps

1. Lint — check R/ directory files for tidyverse style issues:

   Rscript scripts/agents/tidyverse_check.R
   

2. Geospatial review — invoke the geo agent on data/ and any spatial processing scripts that changed since the last commit.
3. Statistical checks — invoke the stats agent on any modelling or aggregation scripts that changed since the last commit.
4. GitHub Actions compatibility — invoke the gh_actions agent on _targets.R and any changed R/ scripts.
5. Targets pipeline — invoke the targets agent on _targets.R and R/tar_release_storage.R.

Output

Print a consolidated report:

[PASS|WARN|FAIL]  lint
[PASS|WARN|FAIL]  geo
[PASS|WARN|FAIL]  stats
[PASS|WARN|FAIL]  gh_actions
[PASS|WARN|FAIL]  targets

Exit non-zero if any check returns FAIL.

Notes

• Skip checks that have no relevant changed files (report as SKIP).
• Be lightweight — do not re-run expensive checks on unchanged files.