
garrytan/gbrain
51 skills7.8k installs1M starsGitHub
Install
npx skills add https://github.com/garrytan/gbrainSkills in this repo
1Brain OpsCovers operational playbooks for Garry Tan’s gbrain knowledge system: how to capture ideas, maintain note structure, execute recurring workflows, and iteratively tune personal productivity rails so the brain stays current and actionable over time.293installs2Idea IngestThe idea-ingest skill for garrytan/gbrain normalizes and stores incoming ideas, snippets, and references into the personal knowledge graph so Claude Code agents can search, relate, and reuse them during ideation and planning.237installs3Queryquery exposes garrytan/gbrain retrieval to agents and users: natural-language questions run against the indexed knowledge graph and media corpus, returning grounded excerpts and summaries for ongoing research, writing, and operational decisions.235installs4MaintainThe maintain skill for garrytan/gbrain provides operational procedures to keep the knowledge base healthy: validating stores, refreshing indexes, and applying routine fixes so agents continue retrieving accurate, current context.227installs5EnrichEnriches entries in the garrytan gbrain personal knowledge system with metadata, cross-links, and summaries. Helps agents keep notes structured and retrieval-ready so research artifacts compound into a searchable, automation-friendly second brain over time.219installs6Meeting IngestionThe meeting-ingestion skill for garrytan/gbrain parses meeting recordings and notes, structures attendees, decisions, and tasks, and writes them into the brain so Claude Code can reference real conversations during research and planning.213installs7SetupThe setup skill from garrytan/gbrain walks through first-time project initialization for the gbrain agent framework. It configures dependencies, environment, and project structure so developers can start building agent workflows immediately instead of debugging missing config or inconsistent local layouts.186installs8Data ResearchGbrain skill for data research across an ingested knowledge graph: formulate questions, pull relevant sources and metrics, and return evidence-backed conclusions. Targets operators validating ideas with structured analysis instead of ad-hoc web searching.182installs9Daily Task ManagerThe daily-task-manager skill operationalizes day-to-day execution in gbrain by maintaining live task lists, priorities, and completion state so agents and users share one actionable view of what must happen today.179installs10Skillifyskillify transforms existing garrytan/gbrain prompts, shell scripts, and procedural runbooks into Claude Code-compatible skills with clear triggers and steps, letting teams promote one-off automations into documented, invocable agent tooling.177installs11Citation FixerThe citation-fixer skill audits and corrects reference formatting across gbrain knowledge outputs, ensuring claims link to real sources with consistent citation style and working URLs.176installs12Media Ingestmedia-ingest adds multimodal intake to garrytan/gbrain: images, audio, video, and documents are parsed, normalized, and stored so agents can search, cite, and reason over ingested assets instead of ephemeral chat context.176installs13Cron SchedulerThe cron-scheduler skill from garrytan/gbrain helps you configure and maintain scheduled tasks for agent workflows. It standardizes how recurring jobs are defined, triggered, and kept running so gbrain automations do not depend on ad hoc timers or one-off scripts scattered across the codebase.175installs14Soul Auditsoul-audit reviews garrytan/gbrain agent soul definitions—persona, values, tone, refusal boundaries, and tool permissions—for contradictions, safety gaps, and drift before agents ship to production, shared workspaces, or automated publish-and-query loops.175installs15BriefingGbrain skill that produces concise briefings from a personal or team knowledge base: ingested articles, notes, and data are summarized into decision-ready updates. Useful during early research when operators need fast context without re-reading every source.173installs16Cross Modal ReviewThe cross-modal-review skill performs multimodal QA on gbrain agent deliverables, verifying that visuals, transcripts, and narrative text agree so mixed-media knowledge entries ship with consistent, auditable meaning.173installs17Daily Task PrepThe daily-task-prep skill runs end-of-day or morning planning in gbrain, synthesizing context into a prioritized agenda and pre-staged agent actions so daily execution starts with clarity instead of rediscovery.173installs18IngestGbrain skill for ingesting external content—RSS, documents, APIs, and notes—into a unified searchable store with normalization and refresh workflows. Foundational build step that feeds briefing and data-research skills with up-to-date corpora.173installs19Repo ArchitectureThe repo-architecture skill builds a durable mental model of the garrytan/gbrain codebase for agents, documenting layers, dependencies, and conventions so subsequent coding and skill work stays consistent with project structure.172installs20Signal DetectorSignal-detector from garrytan/gbrain helps agents watch heterogeneous inputs and flag emergent patterns—market moves, community chatter, or technical shifts—so founders prioritize hypotheses with evidence instead of intuition alone.172installs21MigrateThe migrate skill for garrytan/gbrain guides controlled upgrades of knowledge stores and configuration across versions or hosts, ensuring embeddings, metadata, and cross-references survive operational changes.169installs22Reportsreports compiles garrytan/gbrain ingested media, queries, and signals into structured documents—weekly digests, research memos, and executive briefs—so teams get repeatable narrative outputs without hand-assembling sources each cycle.168installs23TestingThe testing skill executes and authors validation for garrytan/gbrain, covering skill handlers, scripts, and integrations so agent automation changes are verified through automated checks before they reach daily production use.168installs24Skill Creatorskill-creator guides authoring of new Claude Code skills for garrytan/gbrain: defines triggers, SKILL.md structure, references, and helper scripts so repeatable agent behaviors—ingest, query, publish—ship as installable skills instead of ad-hoc instructions.167installs25Publishpublish turns gbrain artifacts—reports, summaries, and detected signals—into channel-ready posts or documents and pushes them to external distribution endpoints, closing the loop from private knowledge work to public or team-facing content.165installs26Webhook TransformsThe webhook-transforms skill for garrytan/gbrain implements transformation rules that convert external webhook JSON into canonical gbrain events and entities, enabling automated ingestion from SaaS tools and custom HTTP sources.164installs27Minion OrchestratorThe minion-orchestrator skill for garrytan/gbrain implements a coordinator pattern that routes work to focused sub-agents, tracks their outputs, and synthesizes answers—core multi-agent tooling for the gbrain Claude Code stack.162installs28Skillpack CheckThe skillpack-check skill for garrytan/gbrain reviews agent skill packages for valid metadata, required sections, compatible dependencies, and naming conventions so broken packs are caught before deployment to Claude Code.159installs29Frontmatter GuardA developer tool for AI integration and automation. This is a developer tool for building and integrating AI-powered features.152installs30Smoke TestA developer tool for AI integration and automation. This is a developer tool for building and integrating AI-powered features.146installs31Book MirrorProduces a personalized book analysis where the left column preserves chapter content and the right column mirrors it to the reader's own situations, people, and patterns. A developer uses it to read a book through the lens of their own life, with optional PDF output.139installs32Strategic ReadingMaps a source text onto a strategic problem the reader is facing and extracts an actionable playbook rather than a general summary. A developer uses it to pull directly applicable guidance from a source for a real decision.134installs33Academic VerifyVerifies research claims by routing through perplexity-research and formatting a citation-checked knowledge-base page with a verdict. A developer uses it when a book, article, or conversation cites a study they want to confirm.131installs34Concept SynthesisMerges duplicate concept stubs, scores and tiers them from Canon to Riff, and synthesizes evolution narratives for the top tiers. A developer uses it to turn scattered idea notes into a curated intellectual fingerprint.129installs35Brain PdfGenerates a clean PDF from any knowledge-base page using the gstack make-pdf binary, treating the page as the source of truth. A developer uses it to deliver a book mirror, playbook, or briefing as a portable document.128installs36Archive CrawlerExplores personal archives within an explicit allow-list, filters for the user's own writing and ideas, and ingests them as knowledge-base pages interactively. A developer uses it to mine old files, emails, and letters for durable knowledge.127installs37Article EnrichmentTransforms wall-of-text article pages into quotable, actionable knowledge-base pages while preserving the raw source in a collapsed block. A developer uses it to make ingested articles useful and referenceable.127installs38Perplexity ResearchCombines existing knowledge-base context with Perplexity's web search to surface the delta with citations, writing results to a research page. A developer uses it for entity enrichment, current-state checks, and freshness deltas rather than simple URL fetches.127installs39Voice Note IngestFiles a voice memo as a knowledge-base page with the transcript kept verbatim, routing content to originals, concepts, people, ideas, or personal notes. A developer uses it to capture audio thoughts without paraphrasing away the signal.127installs40Ask UserA reusable choice-gate pattern that presents labeled options with an escape hatch and halts the turn until the user picks. Skill authors use it when a decision point requires human input before proceeding.124installs41Cold StartBootstraps a new brain by walking consent-gated import phases, routing credentials through ClawVisor so the agent never holds raw keys. A user runs it right after setup when they ask 'now what?'118installs42Functional Area ResolverRewrites RESOLVER.md or AGENTS.md so each functional area lists its sub-skills in a dispatcher clause the LLM routes through. A developer uses it when a routing table has grown too large and eats context budget.110installs43Skillpack HarvestThe inverse of scaffold: an editorial workflow that promotes a mature host skill upstream into gbrain so other clients can scaffold it. A developer uses it to share a battle-tested skill after a privacy-lint and conformance pass.100installs44CaptureThe single human-facing ingestion command that captures content to an inbox slug, queryable immediately, with content-hash dedup. A developer uses it for the simple 'I have a thought, save it' case rather than bulk import.96installs45Schema AuthorManages the brain's schema pack: add, remove, and update page types, aliases, prefixes, and link types, and detect candidate types from the corpus. A developer uses it when notes are untyped or a custom domain needs new typed entities.95installs46Brain TaxonomistA filing gate consulted before creating any page, mapping the content's primary subject to a page type and path from the live schema. A developer uses it to prevent misfiling at write time and to detect taxonomy drift.94installs47EiirpThe universal post-work organizer that ensures nothing from a session lives only in chat, filing enriched pages and auditing the skill graph. A developer uses it to make research permanent and work re-doable after any significant session.93installs48Schema UnifyRuns the migration from gbrain-base to gbrain-base-v2, using onboard checks and the unify-types handler to collapse about 94 types to 15 canonical ones. A developer uses it when type proliferation or a pack upgrade is flagged.86installs49Skill OptimizerTreats SKILL.md as trainable parameters and iteratively rewrites the body, gating every candidate on a validation set before accepting. A developer uses it to improve a skill's execution quality against a benchmark or after switching target models.79installs50Gbrain UpgradeTurns the UPGRADE_AVAILABLE marker printed by gbrain commands into the right upgrade action for the operator's chosen mode. A developer relies on it to keep gbrain current without running a command parsed from untrusted output.70installs51Idea LineageFollows a single concept across the knowledge base over time, surfacing where it first appeared, what it displaced, and what version is alive now. A developer uses it to understand how their thinking about one topic changed, not to build a broad concept map.65installs