Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of recent attention, adoption, and active maintenance.
An agent-engineering curriculum with a practical second-brain project attached. The skills belong to a larger learning path through context, execution loops, evaluation, and maintaining knowledge in Markdown.
WHO IT'S FOR
PKM power users
automate a self-organizing Obsidian vault
AI agent builders
implement rigorous agent evaluation and auditing
Knowledge engineers
manage large-scale information ingestion and backfilling
Compatible AgentsThe repository documents support for these agents. The skills may also work with other agents that can load SKILL.md files, but they may need some setup or small changes.
Make sure second-brain-quiz is installed in your agent before trying this task.
Use second-brain-quiz to quiz me on a topic I choose from my vault. Use a handful of relevant concept pages and ask four questions covering recall and application. Withhold answers until I respond, then grade against the pages and identify which pages I should reread.
An article topic supported by material in the vault.
Make sure second-brain-write is installed in your agent before trying this task.
Use second-brain-write to build a short article outline on a topic I choose from my vault. Find the relevant concept pages, suggest an angle, and tie each outline point to its supporting pages. Surface disagreements, weak evidence, and missing material. Stop at the outline.
Source title, channel or speaker, URL, and date for frontmatter.
Existing timestamps and speaker information, if available.
Make sure second-brain-transcript is installed in your agent before trying this task.
Use second-brain-transcript to clean the short raw transcript I provide. Preserve all meaning without summarising: add punctuation and paragraphs, label speaker changes, correct clear technical-term errors, and flag uncertain terms. Retain coarse timestamps where available and add frontmatter using the source metadata I provide. Return the cleaned transcript and a brief list of corrections, uncertainties, and any suggested topic splits.