agentleFS
benororbenoror/obsidianos_work/AGENTS.md

obsidianos_work

AGENTS.md166 starsChanged 5 months agoMITView on GitHub

This is an Obsidian vault with AI agent skills. See USER.md for the vault owner's identity. Agents working in this repo should be aware of the following structure and conventions. Custom skills live in .agents/skills/. Each has a SKILL.md with usage, workflow, and conventions.

  • Reads credentials

The file

We split this AGENTS.md into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If you forked or cloned this repo into a private vault, you can pull structural updates (skills, rules, shared conventions) without …

    line 117

≈ “Common” is an estimate: at least 1 in 5 of the 41,581 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
AGENTS.md

Agents

This is an Obsidian vault with AI agent skills. See USER.md for the vault owner's identity. Agents working in this repo should be aware of the following structure and conventions.

Skills

Custom skills live in .agents/skills/. Each has a SKILL.md with usage, workflow, and conventions.

SkillWhat it doesArguments
meetingCreate a meeting note
├ /meetingPick from today's Google Calendar events
├ /meeting {title}Manual creation with today's datefolder={subfolder}
├ /meeting wrap <path>Wrap up: mail-transcripts → cache → participants → todos → commit
└ /meeting wrap pending [dates]Find & wrap all pending meetingstoday, yesterday, last week, 2026-01-01..2026-02-03
cache-notesFetch & embed AI transcripts as Obsidian callouts
├ /cache-notes <path>Cache a specific meeting file (prompts for URLs if empty)
├ /cache-notes allBatch-cache all uncached meetings
└ /cache-notes refresh <path>Re-fetch and overwrite existing cache
fill-participantsResolve and fill Participants: frontmatter
├ /fill-participants <path>Fill for a specific meeting note
└ /fill-participants allScan and fill all meetings missing participants
followup-todosExtract action items as plain markdown bullets
├ /followup-todos <path>Extract from a specific meeting note
└ /followup-todos/note-status pending --step=todos
recapProduce a recap from emails, Slack, Jira, and vault notes
├ /recapRecap current week (Monday through today)
└ /recap [dates]Recap a specific date rangetoday, yesterday, last week, 2026-01-01..2026-02-03
sync-upstream-obsidianosPull structural updates from upstream ObsidianOS
├ /sync-upstream-obsidianosPreview and merge upstream updates
└ /sync-upstream-obsidianos previewShow what's new without merging
note-statusVerify meeting notes are fully processed
├ /note-status <path>Check a specific note
├ /note-status [dates]Check notes in a date rangetoday, yesterday, last week, 2026-01-01..2026-02-03
├ /note-status allCheck every meeting note
└ /note-status pending [dates]Filter to pending, prompt selection--step=notes|cache|participants|todos
commitStage and commit with flexible intent parsing
├ /commitStaged files, or infer related changes
├ /commit <file or folder>Scope commit to a specific path
├ /commit <description>Infer scope from free-text intent
├ /commit amend [scope/description]Amend last commit (with optional scope or new message)
└ (sequence mode)Deferred — sub-skills skip, caller commits once at the end
defuddleExtract clean markdown from web pages with Defuddle CLI (prefer over raw fetch for articles & docs)
json-canvasCreate and edit Obsidian JSON Canvas (.canvas) files
obsidian-basesCreate and edit Obsidian Bases (.base) — views, filters, formulas
obsidian-cliDrive a running Obsidian app via the obsidian CLI (read/write/search, dev helpers)
obsidian-markdownObsidian Flavored Markdown — wikilinks, embeds, callouts, properties
mail-transcriptsDrain Gmail AI-transcript notifications into meeting Notes: (gws)
organize-meetingsTriage Meetings/_inbox/ into the right subfolder via qmd + memory<path>, --auto
otter-fetchOtter.ai transcript sub-skill (invoked by cache-notes / fill-participants)
people-stubsBatch-create Teams/People/@Name.md stubs from unresolved [[@*]] links--auto
onboard-personCreate/enrich a @Name.md from external sources (ATS / profile)<Name>, --candidate
last-1on1Flag overdue 1:1s by recency (Meetings.base OneOnOnes view)<weeks>, @Name
vault-healthAudit unresolved links, orphans, deadends, sync conflictsunresolved, orphans, deadends, sync, --fix-people
migrate-tagsMigrate body #hashtags to frontmatter tags: lists<path>, --dry-run, --strip-body
workspace-modeSwitch named Obsidian Workspace layouts<name>, save <name>
sprint-retroSynthesize a Sprint Retro draft from vault sources<N>, <dates>

Skill proxies (symlinks to .agents/skills/)

The canonical skill tree is .agents/skills/. These paths are symlinks to the same directory so each agent discovers the same skills:

PathUsed by
.cursor/skillsCursor
.claude/skillsClaude Code
.opencode/skillsOpenCode
skills/OpenClaw (workspace skills)

OpenCode also discovers .agents/skills/ directly; the .opencode/skills link matches the Cursor / Claude layout.

Rules

Shared rules live in .agents/rules/ and are the single source of truth. Agent-specific wrappers point to them.

RulePurpose
qmd-searchPrefer QMD over grep for vault search (always apply)
skill-conventionsSkill structure, shared conventions, project context
skill-registryKeep skill tables in sync when skills change

Agent wrappers

AgentInstruction filesHow rules are loaded
Cursor.cursor/rules/*.mdcAuto-injected by glob; each .mdc points to .agents/rules/
Claude CodeCLAUDE.md, .agents/skills/CLAUDE.mdRead on startup; point to .agents/rules/
OpenCode / CrushOpenCode.mdRead on startup; points to .agents/rules/
OpenClawAGENTS.md (and your gateway config)Rules live in .agents/rules/; skills via repo skills/ → .agents/skills/

Vault Layout

Meetings/              Meeting notes by subfolder (Engineering/, One-on-ones/, Interviews/, etc.)
Templates/             Obsidian templates
Teams/People/          Person files: @Name.md
Teams/                 Team files: +TeamName.md
ToDo's.md              Aggregated Tasks plugin view
Tracker.md             Project tracker

Conventions

  • User: See USER.md for identity, timezone, aliases, and agent behavior rules.
  • Frontmatter: YAML between --- fences. modified: is managed by Obsidian — never set manually.
  • Participants: [[@Name]] wikilinks (people) or [[+Team]] (teams).
  • AI Transcripts: Cached under ## 🤖 AI Notes with provider subheadings and collapsible callouts ([!gemini_notes]-, [!gemini_todos]-, [!gemini_transcript]-).
  • Tasks priorities: 🔺 highest, ⏫ high, 🔼 medium, 🔽 low.

MCP & external tools

  • QMD — Configured in .cursor/mcp.json for vault search (MCP server npx qmd mcp).
  • Google Workspace — Use the gws CLI (Google Workspace CLI) for Docs, Drive, Calendar, and Gmail in read-only workflows. See README.md § Google Workspace CLI and .agents/skills/_shared/google-workspace-cli.md.

Syncing from Upstream

If you forked or cloned this repo into a private vault, you can pull structural updates (skills, rules, shared conventions) without overwriting your personal data.

./.scripts/sync-upstream.sh              # fetch + merge
./.scripts/sync-upstream.sh --preview    # see what's new without merging

First-time setup:

git remote add upstream <url-to-this-repo>
./.scripts/sync-upstream.sh              # auto-configures merge driver on first run

Protected paths (always keep local version during merge): USER.md, Tracker.md, .env, .cursor/mcp.json, Meetings/, Teams/, Templates/, Recaps/. Edit .gitattributes to add or remove paths.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.