AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Obsidian Knowledge Base

skill-spc-jgs-obsidian-kb-skill-qoderwork · by Spc-jgs

Use when the user explicitly asks to save, create, update, archive, or remember content in an Obsidian vault, to file or sort its Inbox, to check or audit the vault's health, or asks which parts of a conversation are worth capturing as durable knowledge. Conversation-harvest analysis and vault audits are read-only, and Inbox filing always shows a plan before it moves anything; never mutate the Va…

No reviews yet
0 installs
0 views
view→install

Install

$ agentstack add skill-spc-jgs-obsidian-kb-skill-qoderwork

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-spc-jgs-obsidian-kb-skill-qoderwork)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
today

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Obsidian Knowledge Base? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Obsidian Personal Knowledge Base

Overview

Agent-agnostic instructions to create, organize, and update notes in an Obsidian vault. Any tool that reads/writes local files can use it.

DO NOT auto-save

This skill never writes to the vault on its own. Write only after explicit save intent ("save to Obsidian", "沉淀", "总结存档"). An explicit request to evaluate what from a conversation is worth capturing may run conversation-harvest.md as analysis only. For unrelated Q&A, debugging, or casual chat: do nothing.

When the user asks to save or review capture candidates

  1. Route analysis before Vault discovery: for an analysis-only conversation review, read conversation-harvest.md, return the candidate proposal, and stop without locating or scanning a Vault. Otherwise continue.
  2. Find vault for a write: env OBSIDIAN_KB_VAULT~/.obsidian-kb-config → ask. Refuse unless a real vault (.obsidian/ + Templates/).
  3. Read the smallest operation-specific reference set before writing. Read Vault governance first, then let one compact discovery call answer this: required_references names every file the selected type, template, and destination require, so read that set instead of discovering it one failure at a time. Pass the governed route as --folder — a crowded child folder does not make its parent look crowded. The full map:
  • New note: read only note-creation.md.
  • Finished source-backed article: after note-creation.md routes it, also read

web-capture.md; verified captures additionally load deep-capture.md.

  • Crowded selected destination: when compact discovery reports it, also read

folder-routing.md; uncrowded destinations do not load it.

  • Existing note: read only update-note.md.
  • Filing the Inbox: read only process-inbox.md. Filing moves notes the user

already wrote; it never authors one, so ≤1 note written does not bound it.

  • Conversation context archive: read conversation-digest.md.
  • Conversation knowledge review: read conversation-harvest.md; load

note-creation.md only after one durable candidate is selected for writing.

  • Task Memory: read task-memory.md only after explicit opt-in (off by default).
  • YAML, rules, and Git references are troubleshooting or post-processing material: load yaml-standards.md, rules-and-errors.md, or git.md only when the current task requires it.
  1. Prefer bundled helpers (never a one-off script): run python /scripts/run_helper.py ..., where ` contains this SKILL.md. For new notes use create-note --preflight-json, inspect the structured validation, then apply with --from-preflight --apply --compact-json instead of resending the body; update-note is only for Task Memory, while ordinary existing-note edits follow update-note.md` with native file tools.
  2. Stay bounded: ≤10 files scanned, ≤1 note written, ≤5 wikilinks. Never overwrite; add -2 on name clash. Validate after.
  3. Route: daily→15-Daily · meeting→10-Work · learning/article→20-Learning · insight/digest→30-Insights · project→40-Projects · person→50-People · quick/unread source→00-Inbox. An ordinary finished article uses standard capture; verified capture is explicit or evidence-sensitive.

When the user asks to check or audit the Vault

Read only audit-vault.md. The audit is read-only and reports findings; it never repairs a note. Fixing one is a separate request needing its own explicit save intent — reporting a finding does not grant authority to act on it.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.