Menu
Simulate a CRAN reviewer evaluating myIO for acceptance. Adversarial review that flags rejection risks before you submit.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Carry one human-approved myIO package issue through test-first implementation, package checks, and an unmerged PR without releasing or submitting to CRAN. Use for scheduled myIO backlog work.
Research myIO package, chart-engine, MCP, CRAN, compatibility, and performance opportunities and create evidence-backed research candidates. Use for scheduled myIO scouting or inbound enhancement triage.
Run R CMD check --as-cran on the myIO package, parse the output, and map each ERROR/WARNING/NOTE to a concrete fix
End-to-end CRAN submission workflow for myIO — pre-checks, fix cycles, tarball build, cran-comments.md, and resubmission handling. Use when preparing to submit or resubmit to CRAN, or when you receive CRAN reviewer feedback.
Bump version (semver-computed from what's actually merged), finalize NEWS.md, run the full CRAN-readiness gate, build the tarball, tag and create the GitHub Release. Stops before the actual CRAN upload/email confirmation — that step needs the maintainer. Designed for scheduled one-time invocation.
Document and enforce R package versioning strategy for myIO. Covers semantic versioning, backward compatibility policy, deprecation schedules, and CRAN constraints. Use when planning a release, deprecating a function, or documenting versioning decisions.
| name | cran-review |
| description | Simulate a CRAN reviewer evaluating myIO for acceptance. Adversarial review that flags rejection risks before you submit. |
Simulate a CRAN volunteer reviewer evaluating myIO for acceptance. Perform the review directly or use a read-only review subagent when explicitly requested.
Read the core package files:
DESCRIPTIONNAMESPACEcran-comments.mdinst/ directory listing.RbuildignoreIf $ARGUMENTS specifies a focus area, weight that area more heavily but still review everything.
If R CMD check output exists in myIO.Rcheck/00check.log, read it. Otherwise note that check hasn't been run.
Delegate to the cran-reviewer agent with all gathered context.
Return the simulated reviewer email and detailed findings.
/cran-submit — catch issues early/cran-review — verify fixes