Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/agricidaniel/claude-obsidian/agents-mdgit clone --depth 1 https://github.com/AgriciDaniel/claude-obsidianWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.01123 | $0.01123 |
| Opus 5 | $0.00562 | $0.00562 |
| Sonnet 5 | $0.00225 | $0.00225 |
| Haiku 4.5 | $0.00112 | $0.00112 |
Grade A, and why
claude-obsidian AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 98 lines — stays where its author put it; the contents beside it link to each section on GitHub.
claude-obsidian: Agent Instructions
claude-obsidian is a local-first Agent Skills package for building source-cited,
compounding Obsidian knowledge bases. It also ships a Claude Code plugin adapter.
The portable workflow is implemented in skills/ and the standard-library
claude_obsidian/ core; host hooks never define knowledge behavior.
Product and vault boundaries
- This repository is the product source. It is not the default user vault.
- A user vault is the directory containing
.claude-obsidian.json,wiki/, and.raw/. Mutable state always belongs there. templates/vault/is the distributable seed. Rootwiki/,.raw/, and.vault-meta/are contributor state and are excluded from public artifacts.- Never derive a user vault from the plugin cache or
${CLAUDE_PLUGIN_ROOT}. - A checkout containing contributor-vault state has no marketplace catalog.
config/public-marketplace.jsonis injected as.claude-plugin/marketplace.jsononly inside the audited release artifact. An extracted distribution-clean artifact may retain that exact manifest and rebuild idempotently. A public default branch must be populated from the clean artifact, never by pushing contributor-vault state.
Resolve a vault in this order: explicit --vault,
CLAUDE_OBSIDIAN_VAULT, nearest .claude-obsidian.json, then an unambiguous
vault at or above the current directory. Fail closed when no vault is selected.
Bootstrap
- Read this file. In a development checkout, also read the host-only root
CLAUDE.mdwhen present; release artifacts intentionally omit it. - Read the selected skill completely.
- Read only the references that skill routes to.
- Resolve the user vault. If
wiki/hot.mdexists, read it silently. - For a new vault, run
python3 scripts/claude-obsidian.py init PATHfirst; apply only after reviewing the dry run. Useadoptfor an existing vault.
Canonical skills
All 15 skills live at skills/<name>/SKILL.md. They use the portable Agent
Skills frontmatter subset: exactly name and description. Do not add mirrored
files under commands/; Claude invokes plugin skills by namespaced names such
as /claude-obsidian:wiki.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 98 lines · 1,123 tokens per session scan A f12fe295445b
claude-obsidian AGENTS.md is an instructions file published in the GitHub repository AgriciDaniel/claude-obsidian (14,464 stars, last pushed 6d ago), licensed MIT. It adds 1,123 tokens to every session, about $0.0056 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
claude-canvas CLAUDE.md
Instructions for AgriciDaniel/claude-canvas, covering claude-canvas — ai-orchestrated visual production, what it does, directory structure, quick commands and canvas location.
llm-brain CLAUDE.md
Instructions for kimsanguine/llm-brain, covering llm-brain — claude code 운영 지침, 가드레일 (절대 위반 금지), 프라이버시 경계 (agent memory os), 명령어 and ingest.
second-brain-starter-kit AGENTS.md
AGENTS.md instructions for tonydzi/second-brain-starter-kit, covering agents.md — working in this repo, what this repo is, layout, how to verify a change and conventions.
commonplace CLAUDE.md
Instructions for briannaworkman/commonplace: This repository holds two vaults. Nothing should be ingested or edited here at the root.
cmds-llm-wiki AGENTS.md
Schema and harness document for the CMDS LLM Wiki vault. Defines the 3-layer architecture (Raw Sources / Wiki / Schema), ingest-query-lint operations, file conventions, and frontmatter standards. This is the single source of truth for LLM behavior in this vault.
cmds-llm-wiki CLAUDE.md
Schema and harness document for the CMDS LLM Wiki vault. Defines the 3-layer architecture (Raw Sources / Wiki / Schema), ingest-query-lint operations, file conventions, and frontmatter standards. This is the single source of truth for LLM behavior in this vault.