alessandrobologna

Research a user-specified topic with current authoritative sources and package the findings as a reusable topic-kb skill (for example, kinesis-kb). Use when a user asks to learn about a topic, build a knowledge-base skill, or convert web research into reusable guidance.

Create a feature branch in a Git worktree, work independently, then merge back and clean up safely.

Summarize `.waylog/history/*.md` into a sanitized per-session cache + a condensed, chronological `.waylog-journal/summary.md` journal.

Ensure Mermaid diagrams in GitHub READMEs reserve bottom-right safe area using init-only config (no CSS, no spacer nodes) and optionally apply handrawn style on request.

Use when changing the Agent Bus Fumadocs website, docs generator, site metadata, GitHub Pages export behavior, or Markdown docs that feed agentbusmcp.com.

Use when an agent needs to collaborate through Agent Bus, coordinate via topic threads, run reviewer or implementer handoffs, recover from AGENT_NAME_IN_USE, or continue a code-review loop in an existing Agent Bus topic.

Process Markdown Mermaid blocks with mermint using README-style output (`<picture>` + light/dark SVGs), preserve source blocks with `--keep-mermaid`, refresh generated SVG assets safely, and run a visual QA loop on rasterized light/dark previews. Use this skill when: 1. User asks to convert Mermaid code blocks in Markdown files 2. User asks to render/update README diagrams like this repo 3. User asks to regenerate `svgs/*-light.svg` and `svgs/*-dark.svg` assets 4. User asks to run idempotent markdown reprocessing with preserved Mermaid source 5. User asks to create or update Mermaid `architecture-beta` diagrams with AWS service icons

Run CI-style preflight QA checks and guide the user through a clean commit/push workflow by mirroring `.github/workflows/ci.yml`. Use when asked to verify repo readiness, run pre-commit/pre-push checks, ensure CI passes locally, or before committing/pushing changes.

Audit Codex session rollout JSONL files for large retained tool outputs that inflate model context, calculate current token usage for this session, correlate token-count events with compactions, and recommend scoped AGENTS.md rules. Use when the user asks to analyze Codex session usage, calculate current token usage, inspect large command output, reduce tool-output bloat, investigate context inflation, estimate compaction cost, or wants rules to reduce future output consumption.