search

Solid

General vault retrieval — answers both "what" and "why" questions across MEMORY.md, memory/, session logs, project trackers, and vault notes. Uses the search tools (lex+vec+hyde) with grep fallback.

AI & Automation 20 stars 2 forks Updated 5 days ago Apache-2.0

Install

View on GitHub

Quality Score: 84/100

Stars 20%
44
Recency 20%
100
Frontmatter 20%
70
Documentation 15%
100
Issue Health 10%
50
License 10%
100
Description 5%
100

Skill Content

# /search — General vault retrieval ## Purpose First-class retrieval skill that answers both **what** and **why** questions across all knowledge layers in the OneBrain vault. Distinct from existing skills: - `/distill` synthesizes a topic into a *new persisted note* (write); /search is read-only - `/recap` promotes session insights to *memory/* (write); /search is read-only - A direct search query gives content matches; /search ranks by question type and surfaces decision chains ## Sources searched (in order of relevance) 1. `[agent_folder]/MEMORY.md` — always-loaded persona/active-projects 2. `[agent_folder]/memory/*.md` — match via MEMORY-INDEX topics 3. `[logs_folder]/session/**/*-session-*.md` — past session logs 4. `[projects_folder]/**/*.md` — project notes including embedded specs/plans/design docs 5. Project tracker decisions log tables (project MOCs under `[projects_folder]/`) 6. Vault notes (`[knowledge_folder]/`, `[resources_folder]/`, `[areas_folder]/`) 7. `[logs_folder]/checkpoint/*.md` — in-session-state recovery (for current-day questions) ## Tools used - **Search tools lex+vec+hyde** if `search.collection` is configured in onebrain.yml (legacy top-level `qmd_collection` still honored) (preferred) — follows the cascade in `skills/startup/SEARCH.md` - **Glob + Grep fallback** per the cascade's fallback triggers, or unconditionally if the search tools are unavailable - **Heuristic question-type detection**: matches `^why\b` (or the agent's bilingual intent...

Details

Author
onebrain-ai
Repository
onebrain-ai/onebrain
Created
4 months ago
Last Updated
5 days ago
Language
Shell
License
Apache-2.0

Similar Skills

Semantically similar based on skill content — not just same category

AI & Automation Solid

vault-search

Use proactively mid-task, on your own initiative, whenever past work might already cover the current step — before re-fixing a recurring or previously-seen bug, touching an unfamiliar subsystem, or making a risky/hard-to-reverse change — not only when asked. Also when the user wants to recall prior context: 'what did we decide about X', 'find everything about Y', 'что мы решили про', 'что мы знаем', 'как мы делали', 'когда последний раз', 'такое уже было', 'вспомни', 'напомни', 'посмотри в памяти', 'найди в памяти', 'как решали', 'где ещё может быть', 'что было с', or similar. Synthesizes across notes with source citations; prefer over generic memory tools when an Obsidian vault is available.

6 Updated yesterday
jojoprison
AI & Automation Solid

cortex-query

Search and retrieve content from the cortex vault — the user's external memory. Use (1) when the user says "查 cortex", "之前有記過", "cortex 裡有沒有", "check my notes", "what did I write about", AND (2) PROACTIVELY — before answering any non-trivial question about ongoing projects, internal company tooling, cortex-vec, plugin development, NAS debugging, DSM workflows, or any topic that resembles prior work the user might have notes on. When the SessionStart hook lists vault topics and the user's request matches one, search cortex BEFORE asking clarifying questions or guessing. Treat the vault as authoritative prior context: searching it cheaply is always better than reinventing an answer the user already wrote down.

7 Updated today
XBlueSky
AI & Automation Listed

memory-search

Searches the project's recorded memory (core-memory-kit) — decisions, conventions, architecture, the reasoning behind choices, and where things live — and returns a curated, cited summary. Fire whenever the answer might be something the project already established in past work, HOWEVER the question is phrased — any prior decision, convention, rationale, or "how/where/why is it this way" question, including oblique or roundabout asks ("why is everything so spread out?", "remind me what we settled on for X", "how come these files are tiny?"). Also fire when a "[core-memory-kit] Memory available" hint appears on the prompt. The examples are illustrative, not a checklist — prefer recalling over re-deriving an answer from the code. The session-start snapshot is a bounded index; this skill reaches the rest. Skip only when the question is purely about uncommitted or just-edited live code that memory cannot know, concerns this conversation only, or the user asked to ignore memory.

4 Updated today
LH8PPL