AGENTS.md and tdd-workflow skill instructions now explicitly explain how to determine the GitHub issue number and warn agents never to guess or invent one (@copilot, #61).
use_agent() template is reconciled with this package’s AGENTS.md file (#59).
Tests now use stbl::expect_pkg_error_snapshot() directly instead of a locally-defined wrapper; the expect_pkg_error_snapshot() definition has been removed from tests/testthat/helper-expectations.R (#50).
use_github_copilot() now writes workflow files byte-for-byte from the templates, preserving ${{ }} GitHub Actions expressions that were previously corrupted by whisker rendering (#44).
use_agent() now accepts an overwrite argument (default FALSE) and errors if AGENTS.md already exists, consistent with other use_*() functions (#36).
All use_*() functions that accept overwrite now default to FALSE (#36).
vignette("pkgskills") now provides a complete “Get Started” guide covering terminology, use_ai() setup, skill descriptions, and issue-writing best practices (#31).
use_ai() sets up the full AI agent suite in a single call, installing AGENTS.md, the GitHub Copilot workflow, and all selected skills (#28).
use_github_copilot() installs a copilot-setup-steps.yml workflow and its companion reusable install action into .github/workflows/ (#25).
use_skill_tdd_workflow() installs the tdd-workflow skill and bootstraps tests/testthat/helper-expectations.R in the target project (#11).
use_skill_document() installs the document skill and, if absent, creates a starter R/aaa-shared_params.R file from a built-in template (#9).
use_skill_create_issue() installs the create-issue skill, fetching repository metadata from GitHub and rendering a tailored skill template into the project (#3).
use_agent() installs a structured AGENTS.md file, populating the repository overview from the project’s DESCRIPTION (#2).
pkgskills 0.0.0.8000
Temporary bullet so R CMD check is ok with NEWS.md before first release.