Skip to content
Pocketto
Standalone skills

Standalone skills

Nine single-purpose Pocketto skills for debugging, hotfixes, onboarding, Education, design systems, research, plan review, and PR recording.

Standalone skills are lighter, single-purpose, and have no pipeline. Reach for them for everyday work; use the pipeline when a feature genuinely needs the ceremony.

SkillWhen to reach for it
pocket-help"What is Pocket?", which skill to use, how the flow works
pocket-initOnboard an existing project: generate a project guide, optionally calibrate Education
pocket-educationLearn by doing: you implement, the agent teaches, reviews, and guides
bug-huntingFix a bug, debug a failure, audit code for hidden bugs
hotfixSmall-to-medium change where the full pipeline is overkill
brand-designDesign system, creative brief, brand identity, UI tokens
structured-researchValidate an explicit assumption before it enters planning
validate-planReview plans against DRY, YAGNI, TDD, and codebase context
create-prOpen the phase PR linked to the Pocket issue (Enterprise mode)

How they relate to the pipeline

  • hotfix ↔ pocket-grinding is bidirectional. Grinding's triviality off-ramp recommends hotfix for small clear tasks; hotfix escalates to grinding when work turns out architectural or contested.
  • validate-plan is hotfix's review engine. The subagent reviewer that approves hotfix brief-plans validates them with this skill.
  • structured-research feeds planning. A Confirmed verdict can be carried into grinding/planning as a validated input.
  • create-pr is Enterprise-only. It requires Enterprise mode, a successful preflight, and a linked issue.
  • Education changes who implements. In an Education repo, hotfix, bug-hunting, and pocket-development ask before implementing. An explicit handover is always required to switch.

Invocation

On Claude Code: /pocketto:<skill-name> (for example /pocketto:hotfix "bump the rate-limit window to 60s"). On Pi: trigger phrases and skill names, as listed on each page.

On this page