New: 98,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

Scope Mount

by Abilityai in Abilityai/cornelius

Shared read-scope contract - which scopes a knowledge-base read mounts, decided by the read's ROLE (voice / reasoning / lookup / perception-ingestion-maintenance), plus question-triggered reference…

MITAuto-check: notesKnowledge Management

Install Scope Mount

skills CLI
$ npx skills add Abilityai/cornelius --skill scope-mount -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install Abilityai/cornelius scope-mount --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/Abilityai/cornelius.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.claude/skills/scope-mount .claude/skills/scope-mount && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
scope-mount
GitHub stars
109
Token cost
~2.8k tokens
SKILL.md length
1,191 words
Files
2
Skills in repo
54
Repo updated
First seen
Licence
MIT

At a glance

Shared read-scope contract - which scopes a knowledge-base read mounts, decided by the read's ROLE (voice / reasoning / lookup / perception-ingestion-maintenance), plus question-triggered reference…

  • Tasks that involve Knowledge bases
  • SKILL.md covers Purpose, State Dependencies, The four read roles and Conditional reference mounts…, plus 6 more sections
  • Runs Shell scripts from its folder

What it does

Scope Mount is an agent skill from Abilityai/cornelius. Shared read-scope contract - which scopes a knowledge-base read mounts, decided by the read's ROLE (voice / reasoning / lookup / perception-ingestion-maintenance), plus question-triggered reference mounts (Company, Canon, Thinkers, one book), the never-mount list, the core-first/wide-second retrieval shape and the scope-grouped output format. Every read skill references this file and never copies it. Invoke directly to resolve the mount for a question.

Its SKILL.md is about 2.8k tokens, which your agent loads only when the skill is triggered. The skill folder holds 1 other file (for example `resolve_mount.sh`).

It sits in Knowledge Management, covering Knowledge bases. The repository describes itself as: AI-powered second brain template for Claude Code + Obsidian. The licence is MIT.

When your agent uses it

  • Tasks that involve Knowledge bases

Example prompts

  • “/scope-mount”

Requirements

  • A Bash shell
  • Pre-approved tools (allowed-tools): Read, Grep, Glob, Bash

What it can do on your machine

Read from SKILL.md and the folder at commit b9bea90. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves these tools, so the agent can use them without asking each time:

    • Read
    • Grep
    • Glob
    • Bash

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Ships script files (Shell), which the agent can run.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Scope Mount loads about 2.8k tokens when it runs. Until then it costs about 117 tokens; SKILL.md has 1,191 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~117
When it runs · the whole SKILL.md, loaded when a task matches
~2.8k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check: notes

The automated check noted patterns worth knowing about, such as sudo or a known installer.

  • NotePre-approves every shell command (allowed-tools: Bash)SKILL.md
    allowed-tools: Read, Grep, Glob, Bash

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from Abilityai/cornelius at commit b9bea90, republished under its MIT licence (© Abilityai). 1,191 words, ~2,801 tokens.

Download SKILL.mdSave it as .claude/skills/scope-mount/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
scope-mount
description
Shared read-scope contract - which scopes a knowledge-base read mounts, decided by the read's ROLE (voice / reasoning / lookup / perception-ingestion-maintenance), plus question-triggered reference mounts (Company, Canon, Thinkers, one book), the never-mount list, the core-first/wide-second retrieval shape and the scope-grouped output format. Every read skill references this file and never copies it. Invoke directly to resolve the mount for a question.
allowed-tools
Read, Grep, Glob, Bash
automation
autonomous
user-invocable
true
argument-hint
[role=voice|reasoning|lookup] [the question or topic, for the conditional reference mounts]
metadata.version
1.0
metadata.created
2026-09-04
metadata.updated
2026-09-04
metadata.author
Ability.ai
metadata.changelog
1.0: Initial contract - four read roles with standing mounts (voice=core; reasoning and lookup=core,Books,document-insights), question-triggered reference…

Scope Mount

ℹ️ First, set expectations: print one line with this skill's version and its most recent change - the top of metadata.changelog - e.g. scope-mount v1.0 — recent: initial four-role contract. Then proceed. (Skip the banner when applied silently inside a consuming skill.)

Purpose

The scope primitive (ARCHITECTURE.md → "Scope") was built for hygiene: keep the core fingerprint and q-learning clean of encountered material. Its engine default is fail-closed - unset BRAIN_READ_SCOPE → core. That default is correct for the engine and wrong as a retrieval policy, and until 2026-09-04 it was the retrieval policy for every everyday read: /recall, /advise, /decide and eight other skills carried no mount and so read 1,187 of ~6,800 notes. Everything the user has read (Books, Document Insights) and every reference record was invisible to advice and recall. Nobody chose that.

This file is the policy of record for what a read mounts (ARCHITECTURE.md → "Scope Mount Policy"). The engine keeps failing closed; the mount is declared in the playbook, per read, by role.

Contract rule (the reasoning-checks / SOURCE-AUTHORITY.md pattern): this file is the single source of truth. Consuming skills reference it and put the mount inline on their search commands; they never copy the tables. Change a mount here, not in consumers.

State Dependencies

SourceLocationReadWrite
Scope registryresources/local-brain-search/memory_config.py → SCOPE_REGISTRY (tokens: core, books, document-insights, company, canon, thinkers, meta, inbox, output; families accept Books/<slug>, Company/<sub>)✓
Company record titlesBrain/Company/**/*.md✓
Thinker rosterBrain/Thinkers/THINKERS-INDEX.md✓
Book shelfBrain/Books/<slug>/✓
Similarity contractresources/local-brain-search/SIMILARITY-CALIBRATION.md✓

Never writes. Never changes the engine default.

The four read roles

RoleThe read is forStanding mountConsumers
voicethe user's fingerprint IS the product; the answer must be attributable to himBRAIN_READ_SCOPE=core (declared explicitly, never inherited)/get-perspective-on, /create-article (source pass), /synthesize-insights in "my thoughts on X" mode
reasoningthe best-grounded answer, every source attributedBRAIN_READ_SCOPE=core,Books,document-insights - the reasoning mount/advise, /decide, /think-about-it, /dialectic (monk grounding), /abstraction-ladder (rungs), /synthesize-insights (pattern mode), /talk → thinking-partner, incubation-loop evidence pass
lookup"what do I have on X"the reasoning mount, output grouped by scope/recall, /search-vault, /quick-search, /find-connections (anchor + neighbourhood)
perception / ingestion / maintenancesee everything new · dedup against the write target · measure the fingerprintunchanged and already declared: wide (core,Books,document-insights,meta,inbox,output) / the write target / coredomain-watch, auto-discovery, worldview-refresh, integrate-recent-notes, the extractors, /ingest-source, analyze-kb, coherence, lifecycle

Lookup semantics. "What do I think about X" is answered by the core group; "what do I know about X" by all groups. The user narrows on demand: --scope core (or "only my notes") pins the read to core; --scope <token> mounts exactly that.

/talk is reasoning, not voice (the user, 2026-09-04): the thinking partner speaks from the whole knowledge base, including what the user has read; it still labels what it cites.

Conditional reference mounts (reasoning + lookup only; never standing)

Trigger in the questionAddDetect
a person / org / product / deal namecompanya Company record title appears in the question
a named external thinkerthinkersa roster name from THINKERS-INDEX.md appears
a book title or author already on the shelfan extra precision pass at core,Books/<slug> - the wide pass keeps the whole shelfa Brain/Books/ slug matches
direction / strategy / "should we" about the orgcanon - but the route is /canon-advise; inside /advise mount Canon only to surface disagreement as dissent, never as directionthe canon-advise trigger vocabulary

Detection is a cheap title match, run once before the searches (no LLM call, ~50 ms) by the resolver shipped with this skill:

bash
.claude/skills/scope-mount/resolve_mount.sh <voice|reasoning|lookup> "<the question>"
# BRAIN_READ_SCOPE=core,Books,document-insights,company      <- use verbatim on the wide pass
# # trigger company: Trinity
# # extra pass: BRAIN_READ_SCOPE=core,Books/black-swan-taleb  <- a book hit adds a THIRD, precision pass

Rules (whole-word matching; a false positive costs one slightly wider read, nothing else): a Company record title appears whole in the question (single-word titles such as Trinity match as words - agent in a question never mounts BDR agent); any Thinkers name token of 5+ characters appears (the surname is the natural trigger); a Books slug has one distinctive token of 6+ characters or two tokens present, with a stoplist that keeps generic words (thinking, guide, science, agent, ...) from mounting anything. voice always answers core and runs no detector.

A hit adds the token to the mount for that read only (a book hit adds a pass instead). No hit, no mount. A question that is about the org's direction leaves /advise for /canon-advise (dispatch rule in CLAUDE.md).

Show full SKILL.md (502 more words)Show less

Never in a reasoning or lookup mount

ScopeWhy
05-Metaoperational exhaust - changelogs are 88% of the residual cosine-1.000 edges; thinking/watch files are read by grep where a skill needs them (incubation-loop does)
00-Inboxunprocessed captures
04-Outputderivative of core - it would return the user's own articles as evidence for the user's views
DemoCRMdemo data
Company, Canon, Thinkers as standing mountsreference records are facts, not thinking; they enter only on a trigger

These stay perception-only (domain-watch, auto-discovery, integrate-recent-notes already mount them for that purpose).

Retrieval shape: core-first, wide-second

Document Insights alone is 2.3x core, so one wide search with a small --limit can push core out of the top-k. Run two passes in the same parallel batch - core as the spine, wide as the evidence layer:

bash
# one query, two populations, one parallel batch
BRAIN_READ_SCOPE=core                          resources/local-brain-search/run_search.sh "<term>" --limit 3 --json
BRAIN_READ_SCOPE=core,Books,document-insights  resources/local-brain-search/run_search.sh "<term>" --limit 5 --json

Neighbourhood calls follow the same mount as the search that found the anchor (run_connections.sh "<note>" at the reasoning mount reaches Books/DI neighbours). --hubs / --bridges / --stats stay core always - they are the fingerprint, not a read.

Autonomous loops keep --no-track (learning hygiene is about who is reading, not what is mounted).

Output: group by scope, label by provenance

Scope is the note_id folder prefix - no engine change:

GroupPrefixPresent as
your thinking02-Permanent/ 03-MOCs/ AI Extracted Notes/ 01-Sources/core - the fingerprint
what you have readBooks/ Document Insights/encountered, not endorsed - never phrased as "your view"
recordsCompany/ Canon/ Thinkers/facts (provenance: reference), not thinking

Load-bearing sources in a conclusion additionally pass reasoning-checks Check 3 - Provenance Weighting, already wired into /advise, /decide and both crystallization paths. Under the reasoning mount that check is the load-bearing voice guard: it is what keeps a wider read from becoming a wider voice.

What the mount does NOT change

  • Learning. Since Phase 10a (2026-09-04) the learn-gate is per note (learning.py → trains()): the core hits of a mixed read train q-values and get the q-boost, non-core hits never do. Mounting never switches learning off. (Before 10a any mixed read disabled learning - the "landmine" that kept every read at core.)
  • The fingerprint. Hubs, bridges, lifecycle, redundancy: core, by construction.
  • The engine default. Unset scope → core, never "all". The policy lives here and in the playbooks.
  • The endorsement boundary. Reading wider changes what the agent can see, never what it may endorse. encountered → endorsed remains the human act.
  • Thresholds are population-bound. The reasoning mount is its own population; its row lives in SIMILARITY-CALIBRATION.md (Space 2, scoped). Do not carry a core band to the mount or back.

Direct invocation

/scope-mount role=<voice|reasoning|lookup> <question> → runs resolve_mount.sh with those arguments and prints its output: the mount line to use verbatim, the triggers that fired, and any extra precision pass. Useful when writing or debugging a consumer.

Consumers (who declares which role)

/advise · /decide · /think-about-it · /dialectic · /abstraction-ladder · /synthesize-insights · /talk (via thinking-partner) · incubation-loop → reasoning. /recall · /search-vault · /quick-search · /find-connections (+ connection-finder when called from them) → lookup. /get-perspective-on · /create-article → voice (explicit core). Perception, ingestion and maintenance skills keep their existing declared mounts. A consumer with a bare (unmounted) search command is a defect: diagnostics check D3.6 lints for it.

© Abilityai, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 1 other file in .claude/skills/scope-mount of Abilityai/cornelius.

  • SKILL.md
  • resolve_mount.sh

Open the folder on GitHubat commit b9bea90

Compare with similar skills

Scope Mount next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Scope Mount compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Scope Mount this skillAbilityai/cornelius109—~2.8kAutomated safety check: NotesMIT
Knowledge Searchdataelement/bisheng12k—~1.1kAutomated safety check: PassApache-2.0
Capture Conversationoutline/outline41k—~474Automated safety check: PassCustom licence
Find And Citeoutline/outline41k—~537Automated safety check: PassCustom licence
Xhs Virtual Productchenjin-cmd/xhs-virtual-product729—~862Automated safety check: PassMIT
OpenkbVectifyAI/OpenKB4.8k1 repos~2kAutomated safety check: WarnApache-2.0

Similar skills

  • Knowledge Search

    dataelement/bisheng

    Search the user's knowledge bases and knowledge spaces (企业知识库检索).

    12k GitHub stars~1.1k tokensUpdated today
    Knowledge ManagementAuto-check passed

  • Capture Conversation

    outline/outline

    Save the current conversation, a decision, or a set of notes as a document in an Outline collection; use when the user wants to keep what was discussed in their knowledge base.

    41k GitHub stars~474 tokensUpdated today
    Knowledge ManagementAuto-check passed

  • Find And Cite

    outline/outline

    Answer questions from the Outline knowledge base with quotes and links to the source documents; use when the user asks what the team knows, documented, or decided about a topic.

    41k GitHub stars~537 tokensUpdated today
    Knowledge ManagementAuto-check passed

  • Xhs Virtual Product

    chenjin-cmd/xhs-virtual-product

    This skill helps plan, select, produce, and market Xiaohongshu (RED) virtual/digital products — templates, knowledge bases, test tools, study materials.

    729 GitHub stars~862 tokensUpdated 9 days ago
    Knowledge ManagementAuto-check passed

  • Openkb

    VectifyAI/OpenKB

    A skill your agent uses when the user asks about content in their OpenKB knowledge base — research topics, concepts compiled from their documents, cross-document synthesis — or mentions openkb, an…

    4.8k GitHub starsUsed in 1 repo~2k tokens
    Knowledge ManagementAuto-check: warnings

  • Learn From Materials

    dmoshehun-prog/learn-from-materials

    Turns books, PDFs, slides and web pages into a source-grounded knowledge base and an interactive learning page in English or Chinese, with quizzes, relationship maps and reusable methodology notes.

    947 GitHub stars~7.9k tokensUpdated 3 days ago
    Knowledge ManagementAuto-check passed

More from Abilityai/cornelius

All 54 skills in this repo

  • Nano Banana Image Generator

    Abilityai/cornelius

    Generate images using Google's Nano Banana (Gemini 2.5 Flash Image).

    109 GitHub stars~1.2k tokensUpdated 2 days ago
    Auto-check: notes

  • Benchmark Memory

    Abilityai/cornelius

    Systematic benchmarking framework for Local Brain Search memory system with LLM-as-judge scoring

    109 GitHub stars~3.1k tokensUpdated 2 days ago
    Auto-check: notes

  • Changelog Protocol

    Abilityai/cornelius

    Protocol for creating dated changelog files after significant agent sessions.

    109 GitHub stars~555 tokensUpdated 2 days ago
    Auto-check passed

  • Create Article

    Abilityai/cornelius

    Create long-form articles from knowledge base insights. An agent skill from Abilityai/cornelius.

    109 GitHub stars~3k tokensUpdated 2 days ago
    Auto-check: notes

  • Decide

    Abilityai/cornelius

    Structure a decision, not just advise on it. An agent skill from Abilityai/cornelius.

    109 GitHub stars~2.7k tokensUpdated 2 days ago
    Auto-check: notes

  • Epistemic Classification

    Abilityai/cornelius

    Framework for distinguishing research findings from hypotheses and speculative synthesis.

    109 GitHub stars~1.5k tokensUpdated 2 days ago
    Auto-check passed

Categories

Questions about Scope Mount

What does Scope Mount do?

Shared read-scope contract - which scopes a knowledge-base read mounts, decided by the read's ROLE (voice / reasoning / lookup / perception-ingestion-maintenance), plus question-triggered reference…. Scope Mount is an agent skill from Abilityai/cornelius. Shared read-scope contract - which scopes a knowledge-base read mounts, decided by the read's ROLE (voice / reasoning / lookup / perception-ingestion-maintenance), plus question-triggered reference mounts (Company, Canon, Thinkers, one book), the never-mount list, the core-first/wide-second retrieval shape and the scope-grouped output format.

When should I use Scope Mount?

Scope Mount fits situations like: tasks that involve Knowledge bases.

How do I install Scope Mount in Claude Code?

Run `npx skills add Abilityai/cornelius --skill scope-mount -a claude-code`. Or copy the skill folder (.claude/skills/scope-mount in Abilityai/cornelius) into .claude/skills/scope-mount in your project. Claude Code loads it when a task matches its description.

How do I install Scope Mount in Codex?

Run `npx skills add Abilityai/cornelius --skill scope-mount -a codex`. Or copy the skill folder (.claude/skills/scope-mount in Abilityai/cornelius) into .agents/skills/scope-mount in your project. Codex loads it when a task matches its description.

Can I use Scope Mount in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add Abilityai/cornelius --skill scope-mount -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/scope-mount, .gemini/skills/scope-mount, .github/skills/scope-mount and .opencode/skills/scope-mount in your project.

What does Scope Mount need to run?

Going by SKILL.md and its folder, Scope Mount needs a shell for the scripts in its folder. Our summary lists: A Bash shell. Its frontmatter pre-approves these tools: Read, Grep, Glob, Bash.

Does Scope Mount access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Scope Mount safe to install?

Our automated static check of SKILL.md found notes only (pre-approves every shell command (allowed-tools: bash)), nothing it rates as a warning. It is not a guarantee. Review the folder before installing.

What licence does Scope Mount use?

Scope Mount is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Scope Mount use?

About 2.8k tokens (SKILL.md is roughly 11k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Scope Mount?

Skills that share tags, products or a category with Scope Mount: Knowledge Search (dataelement/bisheng, 12k stars), Capture Conversation (outline/outline, 41k stars), Find And Cite (outline/outline, 41k stars) and Xhs Virtual Product (chenjin-cmd/xhs-virtual-product, 729 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Scope Mount?

Abilityai (a GitHub organization) maintains it in Abilityai/cornelius, which has 109 GitHub stars. The repository holds 54 skills in this directory. The repository was last updated on October 8, 2026.

Source: Abilityai/cornelius on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.