Generalv1.1.4

pgc

Helps developers draft their Personio quarterly check-in self-reflection using evidence gathered from their own GitHub, Jira, Slack, Calendar, Confluence, Drive, and Personio peer feedback

by Martin LechnerView on GitHub ↗
performance-reviewself-evalquarterly-check-incareer-growthreflection

pgc

Helps developers draft their Personio quarterly check-in self-reflection using evidence gathered from their own GitHub, Jira, Slack, Calendar, Confluence, Drive, and Personio peer feedback.

Installation

claude plugin install pgc@personio-claude-code-marketplace

What does this plugin do?

Personio's quarterly check-ins ask the same 5 questions every quarter. Reconstructing a quarter's worth of work from memory is slow and produces thin answers. This plugin pulls together your own GitHub PRs, resolved Jira tickets, notable Slack threads, talks/demos from your calendar, docs you've written, and peer feedback you've received, then walks you through the 5 questions with that evidence in hand — you provide the judgment, the plugin provides the paper trail.

pgc-quarterly

Draft your Personio quarterly check-in self-reflection with evidence gathered from your own activity.

What you need

  • gh CLI authenticated (gh auth status)
  • The claude.ai Personio connector (plain "Personio" — not "Personio Internal MCP App," which points to a demo tenant), plus Atlassian MCP, Google Drive MCP, Google Calendar MCP, and Slack MCP connected (run /mcp to check/authenticate any that are missing — the skill skips unavailable sources rather than failing)

Quick Start

/pgc-quarterly

Example

Run /pgc-quarterly two days before your check-in. It asks which quarter you're reflecting on (defaulting to the one that just ended), looks up your job title and matches it against Personio's job-expectations sheet, then gathers your merged PRs, resolved tickets, notable Slack posts, talks you gave, and docs you wrote for that quarter. It walks you through the 5 check-in questions one at a time, showing the relevant evidence as a memory-jog for each, and writes a cited draft to pgc/2026-Q2-self-eval.md that you copy into the real form.

Why would you want this?

  • Stops "what did I even do this quarter" paralysis by surfacing your own PRs, tickets, docs, talks, Slack activity, and peer feedback automatically
  • Every claim in the draft is cited — no vague, unverifiable statements
  • Never invents your opinions or growth goals — those come from you, the plugin only drafts the surrounding prose
  • Keeps a running paper trail so the year-end cycle isn't a scramble to reconstruct four quarters at once

How evidence gathering works

Each evidence source is gathered by its own dedicated read-only subagent (pgc-evidence-github, -jira, -slack, -calendar, -docs, -feedback), dispatched in parallel. Each agent has a minimal, source-specific tool set — so the "only ever your own data" guarantee is enforced by what each agent can do, not just by its prompt — a pinned model matched to the work (cheap models for mechanical listing, stronger ones where filtering judgment is needed), and a fixed output contract so results merge cleanly. A source that's disconnected returns a one-line note and the run continues without it. The pgc-evidence-docs agent never reads a document it doesn't own; for docs where your name appears but someone else owns them, it surfaces the link and the skill asks you to confirm.