Agent skill

Tool Discovery

by grandamenium in grandamenium/cortextos

Discover, research, and CONNECT the tools a tool-agnostic personal assistant needs — email, calendar, contacts, meeting notes, messaging, CRM.

MITAuto-check: notesProductivity & Automation

Install Tool Discovery

skills CLI
$ npx skills add grandamenium/cortextos --skill tool-discovery -a claude-code

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

GitHub CLI
$ gh skill install grandamenium/cortextos tool-discovery --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/grandamenium/cortextos.git skills-src && mkdir -p .claude/skills && cp -r skills-src/community/agents/agentic-crm-assistant/.claude/skills/tool-discovery .claude/skills/tool-discovery && 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
tool-discovery
GitHub stars
101
Token cost
~1.3k tokens
SKILL.md length
676 words
Files
1
Skills in repo
55
Repo updated
First seen
Licence
MIT

At a glance

Discover, research, and CONNECT the tools a tool-agnostic personal assistant needs — email, calendar, contacts, meeting notes, messaging, CRM.

  • Works in 2 steps: Detect what is already here → The connect loop (run once per domain)
  • Tasks that involve Meeting notes and agendas
  • SKILL.md covers Step 1 — Detect what is…, Step 2 — The connect loop (run…, Worked examples and Record Results
  • Calls brew and npm

What it does

Tool Discovery is an agent skill from grandamenium/cortextos. Discover, research, and CONNECT the tools a tool-agnostic personal assistant needs — email, calendar, contacts, meeting notes, messaging, CRM. CLI-first, research-driven, verified. Use during setup and whenever a workflow fails because a tool is missing or not authed.

Its SKILL.md is about 1.3k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Productivity & Automation, covering Meeting notes and agendas. The licence is MIT.

When your agent uses it

  • Tasks that involve Meeting notes and agendas

Example prompts

  • “/tool-discovery”

Requirements

  • Node.js

Workflow steps

2 steps, taken from the step headings in SKILL.md.

  1. Detect what is already here
  2. The connect loop (run once per domain)

What it can do on your machine

Read from SKILL.md and the folder at commit 6f93838. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • brew
    • npm

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

  • Network

    No URLs in SKILL.md. Its commands use npm, which can reach the network depending on how they are called.

    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

Tool Discovery loads about 1.3k tokens when it runs. Until then it costs about 71 tokens; SKILL.md has 676 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~71
When it runs · the whole SKILL.md, loaded when a task matches
~1.3k

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.

  • NoteMentions a .env fileSKILL.md:68
    where it goes (tool config, `.env`, or org `secrets.env`) — never in chat.

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 grandamenium/cortextos at commit 6f93838, republished under its MIT licence (© grandamenium). 676 words, ~1,265 tokens.

Download SKILL.mdSave it as .claude/skills/tool-discovery/SKILL.md (or your agent's skills folder).
name
tool-discovery
description
Discover, research, and CONNECT the tools a tool-agnostic personal assistant needs — email, calendar, contacts, meeting notes, messaging, CRM. CLI-first, research-driven, verified. Use during setup and whenever a workflow fails because a tool is missing or not authed.

Tool Discovery + Connect Skill

Two jobs: (1) detect what is already installed and authed, and (2) for every domain the user names a service for, research the right CLI, walk the user through connecting it, and verify it actually works — in the Telegram conversation, on the fly.

Preference order for every domain, always: CLI > official connector / MCP > browser automation. A CLI is scriptable, auditable, and does not depend on a live browser session. Only fall back when the tier above genuinely does not exist.

Never ask for secrets in chat. Hand the user the exact auth command to run themselves (in-session with the ! <command> prefix, or in their own terminal) so the token lands in the tool's own secure store, not the transcript.


Step 1 — Detect what is already here

bash
for cmd in gog gh agent-browser peekaboo sqlite3 jq rg; do
  if command -v "$cmd" >/dev/null; then echo "$cmd: $(command -v "$cmd")"; fi
done
test -f .mcp.json && cat .mcp.json
env | grep -E 'GMAIL|GOOGLE|OUTLOOK|NOTION|ZOOM|FATHOM|FIREFLIES|GRANOLA|HUBSPOT|PIPEDRIVE|AIRTABLE|CRM' | sed 's/=.*/=<configured>/'

Anything already installed AND authed: record it and skip to verify. Everything else goes through the connect loop below.


Step 2 — The connect loop (run once per domain)

Domains the CRM assistant covers: email, calendar, contacts, meeting notes, messaging (e.g. iMessage), external CRM. For each one the user says they use:

2a. Ask what service they use for this domain

"What do you use for meeting notes — Granola, Fathom, Fireflies, Zoom, something else?" One domain at a time. Wait for the answer on Telegram.

2b. Identify the CLI (prefer CLI). Research it if you do not know it.
  • If you already know the CLI (Google -> gog, GitHub -> gh), use it.
  • If you do NOT know the CLI for the named service, research it live before asking the user to do anything:
    • WebSearch: "<service> official CLI", "<service> command line auth".
    • WebFetch the tool's install + auth docs.
    • Confirm the actual command and its auth subcommand (do not guess).
    • If the service is already installed locally, read its real help: <cli> --help, <cli> auth --help.
  • Decide the tier: real CLI found -> CLI. No CLI but an MCP/connector exists -> connector. Neither -> browser automation (agent-browser / headed profile) as the last resort, and say so plainly.
2c. Walk the user through install + auth (in chat, on the fly)

Give exact, copy-pasteable steps grounded in what you just researched:

  • Install: the real install command (e.g. brew install ..., npm i -g ...).
  • Auth: the real auth command, run BY THE USER so no secret hits the chat. Tell them to type it in-session with the ! prefix, e.g.: ! gog login you@example.com (Google opens a browser consent, stores a refresh token locally.)
  • If the flow needs an API key/token, tell them exactly where to create it and where it goes (tool config, .env, or org secrets.env) — never in chat.
Show full SKILL.md (248 more words)Show less
2d. Verify with a REAL action (do not trust presence)

Prove the connection works before moving on, with the lightest real read:

  • Email/Calendar (gog): gog auth status, then a 1-item read (list 1 recent email / today's events).
  • Messaging (iMessage): a bounded read of the local Messages DB or the messaging CLI's status/list, scoped to a safe test.
  • Meeting notes: list the most recent note/transcript.
  • External CRM: fetch 1 record. Report the result in Telegram ("Google connected — I can see your inbox and today's calendar"). If verify fails, debug the auth with the user; do not mark it done.
2e. Record it, then loop

Append to TOOLS.md and move to the next domain.


Worked examples

  • Google (email + calendar + drive): CLI = gog. Install if missing, then ! gog login <email>; verify gog auth status + list today's calendar.
  • iMessage (messaging): macOS local. Prefer a messaging CLI / bounded reads of the local Messages SQLite DB; verify by reading the latest message safely. No outbound sends without the approval rule. If no clean CLI exists, say so and propose the connector/automation fallback.
  • Meeting notes: ask which tool (Granola/Fathom/Fireflies/Zoom). Research that tool's CLI or export path live, walk the user through auth, verify by listing the latest transcript.

Record Results

Append a ## Configured Tools section to TOOLS.md with, per tool:

  • domain + provider
  • command or connector name (the tier used: CLI / connector / browser)
  • account identifier, if safe
  • read/write capability
  • approval requirement
  • verified: yes/no + how (the real action you ran)
  • fallback path

Never write secrets to TOOLS.md.

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

Files

Just SKILL.md in community/agents/agentic-crm-assistant/.claude/skills/tool-discovery of grandamenium/cortextos.

Open the folder on GitHubat commit 6f93838

Compare with similar skills

Tool Discovery 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.

Tool Discovery compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Tool Discovery this skillgrandamenium/cortextos101—~1.3kAutomated safety check: NotesMIT
Customer Call Summaryin-the-weeds-hannah-stulberg/team-os-example-repo173—~3.3kAutomated safety check: PassCustom licence
Meeting Prepgnekt/My-Brain-Is-Full-Crew3.9k—~2.7kAutomated safety check: WarnCustom licence
Gws Workflow Meeting Prepgoogleworkspace/cli31k—~300Automated safety check: PassApache-2.0
Capture Meetingkbanc85/claudia296—~2.1kAutomated safety check: PassCustom licence
Zohosundial-org/awesome-openclaw-skills663—~2.7kAutomated safety check: NotesNone

Similar skills

  • Customer Call Summary

    in-the-weeds-hannah-stulberg/team-os-example-repo

    Customer Call Transcript Processing

    173 GitHub stars~3.3k tokensUpdated 4 mo ago
    Productivity & AutomationAuto-check passed
  • Meeting Prep

    gnekt/My-Brain-Is-Full-Crew

    Prepare a comprehensive brief for an upcoming meeting. An agent skill from gnekt/My-Brain-Is-Full-Crew.

    3.9k GitHub stars~2.7k tokensUpdated 3 mo ago
    Productivity & AutomationAuto-check: warnings
  • Gws Workflow Meeting Prep

    googleworkspace/cli

    Google Workflow: Prepare for your next meeting: agenda, attendees, and linked docs.

    31k GitHub stars~300 tokensUpdated 4 days ago
    Productivity & AutomationAuto-check passed
  • Capture Meeting

    kbanc85/claudia

    Process meeting notes or transcript to extract decisions, commitments, and insights.

    296 GitHub stars~2.1k tokensUpdated 1 mo ago
    Productivity & AutomationAuto-check passed
  • Zoho

    sundial-org/awesome-openclaw-skills

    Interact with Zoho CRM, Projects, and Meeting APIs. An agent skill from sundial-org/awesome-openclaw-skills.

    663 GitHub stars~2.7k tokensUpdated 7 mo ago
    Productivity & AutomationAuto-check: notes
  • Meeting Prep

    techwolf-ai/ai-first-toolkit

    Comprehensive pre-meeting briefing that gathers all relevant context from Slack, email, Google Docs, Notion, and calendar.

    132 GitHub stars~1k tokensUpdated 11 days ago
    Productivity & AutomationAuto-check passed

More from grandamenium/cortextos

All 55 skills in this repo
  • Cortext Self Diagnosis

    grandamenium/cortextos

    Diagnose cortextOS itself when the framework misbehaves — an agent has gone silent or wedged, agents are crash-looping, Telegram or agent-to-agent messages are not arriving, crons did not fire, an…

    101 GitHub stars~3.7k tokensUpdated 18 days ago
    Auto-check passed
  • Activity Channel

    grandamenium/cortextos

    You have completed something significant and want the whole org — all agents and the user — to know about it.

    101 GitHub stars~624 tokensUpdated 18 days ago
    Auto-check passed
  • Agentcard Purchase

    grandamenium/cortextos

    You need to make a purchase on behalf of the user — buy a SaaS subscription, pay for an API, purchase a domain, or any transaction requiring a credit card.

    101 GitHub stars~1.1k tokensUpdated 18 days ago
    Auto-check passed
  • Claude To Codex Migration

    grandamenium/cortextos

    Migrate ANY cortextOS agent from the claude-code runtime to the live codex-app-server runtime.

    101 GitHub stars~12k tokensUpdated 18 days ago
    Auto-check: warnings
  • Bus Reference

    grandamenium/cortextos

    Complete cortextos bus CLI reference - all available commands with examples.

    101 GitHub stars~3.8k tokensUpdated 18 days ago
    Auto-check passed
  • Business News Monitor

    grandamenium/cortextos

    Daily cron-driven scan of news/forums/social in a domain to surface market shifts, new competitors, regulatory changes, and net-new opportunities.

    101 GitHub stars~1.3k tokensUpdated 18 days ago
    Auto-check passed

Questions about Tool Discovery

What does Tool Discovery do?

Discover, research, and CONNECT the tools a tool-agnostic personal assistant needs — email, calendar, contacts, meeting notes, messaging, CRM. Tool Discovery is an agent skill from grandamenium/cortextos. Discover, research, and CONNECT the tools a tool-agnostic personal assistant needs — email, calendar, contacts, meeting notes, messaging, CRM.

When should I use Tool Discovery?

Tool Discovery fits situations like: tasks that involve Meeting notes and agendas.

How do I install Tool Discovery in Claude Code?

Run `npx skills add grandamenium/cortextos --skill tool-discovery -a claude-code`. Or copy the skill folder (community/agents/agentic-crm-assistant/.claude/skills/tool-discovery in grandamenium/cortextos) into .claude/skills/tool-discovery in your project. Claude Code loads it when a task matches its description.

How do I install Tool Discovery in Codex?

Run `npx skills add grandamenium/cortextos --skill tool-discovery -a codex`. Or copy the skill folder (community/agents/agentic-crm-assistant/.claude/skills/tool-discovery in grandamenium/cortextos) into .agents/skills/tool-discovery in your project. Codex loads it when a task matches its description.

Can I use Tool Discovery 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 grandamenium/cortextos --skill tool-discovery -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/tool-discovery, .gemini/skills/tool-discovery, .github/skills/tool-discovery and .opencode/skills/tool-discovery in your project.

What does Tool Discovery need to run?

Going by SKILL.md and its folder, Tool Discovery needs the command-line tools its instructions call (brew and npm). Our summary lists: Node.js.

Does Tool Discovery access the network?

SKILL.md contains no URLs. Its commands use npm, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Tool Discovery safe to install?

Our automated static check of SKILL.md found notes only (mentions a .env file), nothing it rates as a warning. It is not a guarantee. Review the folder before installing.

What licence does Tool Discovery use?

Tool Discovery 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 Tool Discovery use?

About 1.3k tokens (SKILL.md is roughly 5.1k 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 Tool Discovery?

Skills that share tags, products or a category with Tool Discovery: Customer Call Summary (in-the-weeds-hannah-stulberg/team-os-example-repo, 173 stars), Meeting Prep (gnekt/My-Brain-Is-Full-Crew, 3.9k stars), Gws Workflow Meeting Prep (googleworkspace/cli, 31k stars) and Capture Meeting (kbanc85/claudia, 296 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Tool Discovery?

grandamenium (a GitHub user) maintains it in grandamenium/cortextos, which has 101 GitHub stars. The repository holds 55 skills in this directory. The repository was last updated on September 23, 2026.

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