Release v2.22.0
The short version
Section titled “The short version”Two improvements, both about tidiness and reach, neither changing what the skills do:
- No more double entries. Each skill used to show up twice in the
/menu: once under its full name and once as a short command wrapper. The short wrappers are removed, so each capability now appears once, as the skill. - Codex now sees the skills. pm-skills shipped only a Claude manifest, so Codex listed the plugin but reported “No plugin skills.” This release adds a Codex-native manifest, so Codex discovers the skills the same way Claude Code does.
What changed
Section titled “What changed”- Removed: the 63 hand-maintained short command wrappers (for example
/pm-skills:okr-writer). They duplicated skills that already exist under their full names. - Added:
.codex-plugin/plugin.jsonso Codex packages and surfaces the skills. - Kept: the 10
workflow-*orchestrator commands.
What did NOT change
Section titled “What did NOT change”- Every skill keeps its name.
foundation-okr-writer,deliver-prd,discover-journey-map- all unchanged. Nothing was renamed. - What the skills produce, and how you use them, is identical. This release is about removing a redundant layer and adding packaging, not about behavior.
- Skill names and behavior are unchanged. A handful of cross-reference lines inside some skills and samples that pointed at the old short commands are updated to reference skills by name (the portable form across clients). The generated invocation snippet on each skill page updates to
/pm-skills:<name>on Claude Code ($<name>on Codex).
Do I need to do anything?
Section titled “Do I need to do anything?”Almost certainly not. The skills are all still in the /pm-skills: menu under their names. The one case that needs a small update: if you saved a short command (like /pm-skills:okr-writer) in a script or prompt, switch it to the skill’s full name (/pm-skills:foundation-okr-writer). The skill is the same; only the short wrapper is gone.
The rule: the skill’s full name is the short command name with its category prefix restored (define-, deliver-, develop-, discover-, foundation-, iterate-, measure-, tool-, utility-). A few examples:
| Removed short command | Use the skill |
|---|---|
/pm-skills:okr-writer | /pm-skills:foundation-okr-writer |
/pm-skills:prd | /pm-skills:deliver-prd |
/pm-skills:journey-map | /pm-skills:discover-journey-map |
/pm-skills:note-and-vote | /pm-skills:tool-note-and-vote |
The pattern above covers all 63 skills; either way, every skill is right there in the /pm-skills: menu under its full name.
A few removed commands were verb-named helpers for the maintainer sub-agents; they map to their utility skills rather than by prefix:
| Removed command | Use the skill |
|---|---|
/pm-skills:pm-critic | /pm-skills:utility-pm-critic |
/pm-skills:pm-audit-repo | /pm-skills:utility-pm-skill-auditor |
/pm-skills:pm-draft-changelog | /pm-skills:utility-pm-changelog-curator |
/pm-skills:pm-release | /pm-skills:utility-pm-release-conductor |
Why remove the short names instead of making them the real names? That heavier change (renaming all 63 skills to short names) was fully planned and reviewed, but it touches roughly ten times as much of the repo - every doc, the sample library, the page generators - almost all of it to serve a naming preference rather than to fix the duplication. We chose the smaller, safer change that fixes the actual problem now. The short-name idea is parked and may return as its own future release.
Did anything about the skills get worse? No. The skills are byte-for-byte the same; you just invoke them by their (slightly longer) full name, which is also what every other client already used.
Versioning: this is a minor release. It removes a redundant convenience layer and adds a manifest; no skill capability changed.