By 0xKatze
Framework to iteratively grill knowledge into an Obsidian vault and analyse/test it with agents. Bundles the grill-with-obsidian skill and four read-only analyzer agents.
Analyzes an Obsidian vault for internal contradictions and terminology drift — conflicting definitions, the same concept under different names, decisions that contradict each other or the code. Use during the obsidian-grill "Analyze" stage. Read-only; outputs findings, does not edit the vault.
Finds knowledge gaps in an Obsidian vault — concepts referenced but never defined, decisions recorded without rationale, accumulating Open Questions, and thin/stub notes. Use during the obsidian-grill "Analyze" stage. Read-only; outputs a prioritised gap list.
Audits an Obsidian vault for staleness — notes whose claims have drifted from current code/reality, decisions silently superseded, and index/overview counts that no longer match the vault. Use during the obsidian-grill "Analyze" stage. Read-only; outputs a freshness report.
Multi-perspective critique of a plan or a vault area, role-playing four reviewers in one pass — factual reviewer, senior engineer, security/threat reviewer, and redundancy checker. Use during the obsidian-grill "Analyze" stage to stress-test before committing knowledge. Read-only; outputs each persona's findings plus a merged verdict.
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A framework for turning an Obsidian vault into a living, tested knowledge base.
Knowledge isn't dumped once and left to rot. obsidian-grill runs a flywheel: grill new understanding in (human-in-the-loop), test the vault with deterministic checks, analyse it with read-only agents, then iterate — each turn the vault gets sharper, less contradictory, and more in sync with reality.
① GRILL ──▶ ② TEST ──▶ ③ ANALYSE ──▶ ④ ITERATE ──▶ (back to ①)
skill scripts agents agenda
See docs/ITERATION-LOOP.md for the full cycle.
skills/grill-with-obsidian/ the human-in-the-loop grilling skill (capture)
SKILL.md, GLOSSARY-FORMAT.md, DECISION-FORMAT.md
agents/ read-only analyzer sub-agents (Claude Code format)
consistency-reviewer.md coverage-gap-analyst.md
freshness-auditor.md perspective-panel.md
scripts/ deterministic KB "unit tests"
vault_lint.py broken [[links]], orphans, frontmatter, dup names (non-zero exit on issues)
vault_stats.py coverage %, index drift, staleness, tag taxonomy
kb_report.py runs both → meta/kb-report-<date>.md
docs/ ITERATION-LOOP.md, AGENTS.md
cocoindex/ incremental semantic index layer (CocoIndex) +
related_unlinked.py (runs now, sklearn) — finds
related-but-unlinked & near-duplicate notes (P2)
.claude-plugin/ marketplace.json + plugin.json (install as a plugin)
obsidian-grill packages and connects three skills into one loop:
grill-with-docs (mattpocock, MIT) — the relentless one-question-at-a-time
grilling methodology, glossary discipline, and ADR criteria. Adapted here.wiki-query — how the grilling reads the vault for domain awareness.wiki-update — how crystallised knowledge is written back into the vault.The grilling skill is the capture door; the scripts and agents are the test and analysis that decide what the next capture should focus on.
Vault config. All tooling resolves the vault from ~/.obsidian-wiki/config
(OBSIDIAN_VAULT_PATH), or $OBSIDIAN_VAULT_PATH, or --vault PATH. Works with
any Obsidian vault — nothing is hard-coded.
As a plugin (skill + agents):
/plugin marketplace add 0xKatze/obsidian-grill
/plugin install obsidian-grill
Manual:
git clone https://github.com/0xKatze/obsidian-grill
cp -r obsidian-grill/skills/grill-with-obsidian ~/.claude/skills/
cp obsidian-grill/agents/*.md ~/.claude/agents/
# ② TEST — health of the vault right now
python scripts/vault_lint.py
python scripts/vault_stats.py
python scripts/kb_report.py # writes a dated report into <vault>/meta/
# ① GRILL — capture (in Claude Code)
# "use grill-with-obsidian to stress-test this plan"
# ③ ANALYSE — dispatch the agents (see docs/AGENTS.md), in parallel, read-only
Requirements: Python 3.9+ (standard library only). The grilling skill and agents run inside Claude Code (or any agent that supports SKILL.md / sub-agents).
MIT — see LICENSE. Bundles/adapts grill-with-docs (mattpocock, MIT).
npx claudepluginhub 0xkatze/obsidian-grill --plugin obsidian-grillContext window & cost monitoring with ccusage integration for Claude Code
Unity Development Toolkit - Expert agents for scripting/refactoring/optimization, script templates, and Agent Skills for Unity C# development
Complete creative writing suite with 10 specialized agents covering the full writing process: research gathering, character development, story architecture, world-building, dialogue coaching, editing/review, outlining, content strategy, believability auditing, and prose style/voice analysis. Includes genre-specific guides, templates, and quality checklists.
Comprehensive .NET development skills for modern C#, ASP.NET, MAUI, Blazor, Aspire, EF Core, Native AOT, testing, security, performance optimization, CI/CD, and cloud-native applications
Complete collection of battle-tested Claude Code configs from an Anthropic hackathon winner - agents, skills, hooks, and rules evolved over 10+ months of intensive daily use
Comprehensive SEO analysis plugin for Claude Code. 25 sub-skills (21 core + 1 orchestrator + 1 framework + 2 extension mirrors) and 18 sub-agents cover technical SEO, content quality, schema, sitemaps, Core Web Vitals, local SEO, backlinks, AI/GEO, ecommerce, hreflang, SXO, clustering, drift monitoring, and Google APIs. Includes optional MCP extensions, SPA-aware rendering, portability, and hardened SSRF/DNS-rebinding safe fetchers.
Modern R development skills for Claude Code - tidyverse patterns, rlang metaprogramming, Bayesian inference, performance optimization, and more