Talk to another Oracle agent via contacts + threads. An agent skill from Soul-Brews-Studio/arra-oracle-skills-cli.

MITAuto-check passedAgent Workflows

Install Talk To

skills CLI
$ npx skills add Soul-Brews-Studio/arra-oracle-skills-cli --skill talk-to -a claude-code

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

GitHub CLI
$ gh skill install Soul-Brews-Studio/arra-oracle-skills-cli talk-to --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/Soul-Brews-Studio/arra-oracle-skills-cli.git skills-src && mkdir -p .claude/skills && cp -r skills-src/src/skills/talk-to .claude/skills/talk-to && 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
talk-to
GitHub stars
123
Token cost
~2.3k tokens
SKILL.md length
949 words
Files
1
Skills in repo
31
Repo updated
First seen
Licence
MIT

At a glance

Talk to another Oracle agent via contacts + threads. An agent skill from Soul-Brews-Studio/arra-oracle-skills-cli.

  • Works in 3 steps: Check ψ/contacts.json → has maw field? →… → No contacts? → maw ls 2>/dev/null | grep… → Neither? → use MCP thread (async,…
  • User says talk to
  • SKILL.md covers Usage, Mode 0: No arguments, Step 0: Contacts Lookup and Transport Selection, plus 9 more sections
  • Calls jq and bash

What it does

Talk To is an agent skill from Soul-Brews-Studio/arra-oracle-skills-cli. Talk to another Oracle agent via contacts + threads. Use when user says "talk to", "message", "chat with", or wants to communicate with another agent (e.g. "talk to pulse", "message neo"). Do NOT trigger for OracleNet social feed (use /oraclenet), skill management (use /oracle), or family registry (use /oracle-family-scan).

Its SKILL.md is about 2.3k 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 Agent Workflows, covering Skill management. It works with Model Context Protocol. The repository describes itself as: Install Oracle skills to Claude Code, OpenCode, Cursor, and 12+ AI coding agents. The licence is MIT.

When your agent uses it

  • User says talk to
  • Wants to communicate with another agent (e.g
  • OracleNet social feed (use /oraclenet)
  • Skill management (use /oracle)

Example prompts

  • “talk to”
  • “message”
  • “chat with”
  • “/talk-to”

Workflow steps

3 steps, taken from the first numbered list in SKILL.md.

  1. Check ψ/contacts.json → has maw field? → use maw hey
  2. No contacts? → maw ls 2>/dev/null | grep -q "{agent}" fallback → use maw hey
  3. Neither? → use MCP thread (async, persistent)

What it can do on your machine

Read from SKILL.md and the folder at commit 9feea09. 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:

    • jq
    • bash

    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

Talk To loads about 2.3k tokens when it runs. Until then it costs about 83 tokens; SKILL.md has 949 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~83
When it runs · the whole SKILL.md, loaded when a task matches
~2.3k

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 passed

The automated check found no risky patterns in SKILL.md.

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 Soul-Brews-Studio/arra-oracle-skills-cli at commit 9feea09, republished under its MIT licence (© Soul-Brews-Studio). 949 words, ~2,301 tokens.

Download SKILL.mdSave it as .claude/skills/talk-to/SKILL.md (or your agent's skills folder).
name
talk-to
description
Talk to another Oracle agent via contacts + threads. Use when user says "talk to", "message", "chat with", or wants to communicate with another agent (e.g. "talk to pulse", "message neo"). Do NOT trigger for OracleNet social feed (use /oraclenet), skill management (use /oracle), or family registry (use /oracle-family-scan).
argument-hint
<agent-name> [message] [--maw | --thread | --inbox]
hidden
true
metadata.internal
true

/talk-to - Agent Messaging

Send messages to agents via Oracle threads. Each agent has a persistent channel thread.

Messaging consolidation (2026-07-06): /hey, /contacts, /mailbox, and /inbox were archived to the zombie tier — this skill is the one messaging verb now. For real-time tmux pings use maw hey directly (or --maw here). Archived skills remain installable via arra install -s <name>.

Usage

/talk-to arthur "What's your status?"          # one-shot message
/talk-to arthur --new "Hey, starting fresh"    # skip lookup, create new thread
/talk-to arthur loop ask about their work      # autonomous conversation
/talk-to #42 "follow up on this"               # post to thread by ID
/talk-to --list                                # show channels
/talk-to arthur --maw "quick ping"             # force maw transport (real-time tmux)
/talk-to arthur --thread "async question"      # force MCP thread transport
/talk-to arthur --inbox "offline message"      # force inbox transport (file write)

Mode 0: No arguments

If ARGUMENTS is empty, show usage help then run --list.

Step 0: Contacts Lookup

Always read contacts first. This is the source of truth for agent routing.

bash
date "+🕐 %H:%M %Z (%A %d %B %Y)"
CONTACTS="$(pwd)/ψ/contacts.json"
if [ ! -f "$CONTACTS" ]; then
  # Fallback path
  CONTACTS="$(pwd)/.oracle/contacts.json"
fi

if [ -f "$CONTACTS" ]; then
  MAW=$(jq -r ".contacts.\"$AGENT\".maw // empty" "$CONTACTS")
  INBOX=$(jq -r ".contacts.\"$AGENT\".inbox // empty" "$CONTACTS")
  THREAD=$(jq -r ".contacts.\"$AGENT\".thread // empty" "$CONTACTS")
  REPO=$(jq -r ".contacts.\"$AGENT\".repo // empty" "$CONTACTS")
  NOTES=$(jq -r ".contacts.\"$AGENT\".notes // empty" "$CONTACTS")
  FOUND_IN_CONTACTS=true
else
  FOUND_IN_CONTACTS=false
fi

If agent not found in contacts AND not found in maw ls:

I don't know "{agent}".

  /contacts add {agent}    — register transport info
  /talk-to {agent} --thread "message"  — try MCP thread anyway

Offer /contacts add first. If user insists, fall through to MCP thread.


Transport Selection

FlagTransportBest For
(none)auto — detect bestDefault
--mawmaw hey (tmux sendkeys)Real-time, local fleet, low latency
--threadMCP oracle_threadAsync, persistent, cross-machine
--inboxFile write to ψ/inbox/Offline, no maw/MCP needed

Auto-detect logic (when no flag):

  1. Check ψ/contacts.json → has maw field? → use maw hey
  2. No contacts? → maw ls 2>/dev/null | grep -q "{agent}" fallback → use maw hey
  3. Neither? → use MCP thread (async, persistent)
bash
# Auto-detect: contacts-first, maw ls fallback
if [ -n "$MAW" ]; then
  echo "USE_MAW (from contacts: $MAW)"
elif maw ls 2>/dev/null | grep -q "{agent}"; then
  echo "USE_MAW (from maw ls)"
else
  echo "USE_THREAD"
fi
Replying to a message another oracle sent you (#434)

A ping arrives in your context looking like [m5:digger] …. The reflex is to answer in your own transcript — but that text goes to Nat, not to digger, who is sitting in a different pane waiting. Reply with maw hey or the conversation is one-sided and nobody notices for hours.

bash
maw hey <their-oracle-name> 'your reply'     # name, not a pane id

Never reply to a pane id. maw peek %3001 works, so it is tempting to paste that same %3001 into maw hey — it hard-errors (bare target '%3001' not found locally). Pane ids are a peek-only handle. maw ls -v is the source of truth for targets; a bare maw ls | grep finds the row but not the addressable name, which is why the auto-detect above should be treated as a hint, not an answer:

bash
maw ls -v | grep -i <name>      # resolves the real target
maw hey <target> ping --dry-run # confirm it resolves BEFORE sending

--dry-run is the cheap check. It costs nothing and catches a wrong target before the message goes somewhere confusing.

When using --maw:

  1. Compose message from intent
  2. Use contacts maw name if available: maw hey {MAW or agent-oracle} '{message}'
  3. Optionally maw peek {agent} to check response
  4. Confirm: Sent via maw to {agent}

When using --inbox:

  1. Compose message from intent
  2. Check contacts for inbox path — if empty, error: No inbox path for {agent}. Run /contacts show {agent}
  3. Write message file:
    bash
    SELF="$(basename $(pwd) | sed 's/-oracle$//')"
    echo "$MESSAGE" > "$INBOX/$(date +%Y%m%d_%H%M)_from_${SELF}.md"
  4. Confirm: Dropped to {agent}'s inbox

When using --thread (or auto-detected thread): Fall through to Mode 3 (one-shot) below.

Routing

PatternUse
channel:{agent}Persistent per-agent channel
topic:{agent}:{slug}Topic-specific thread (with --topic)
#{id}Direct thread reference by ID

Mode 1: --list

  1. arra_threads() (no status filter)
  2. Filter titles starting with channel: or topic:, exclude closed
  3. Display: channel:arthur (#42) pending — 12 msgs

Mode 2: --new (fast create)

Skip lookup. One MCP call.

  1. Compose message from intent
  2. arra_thread({ title: "channel:{agent}", message, role: "human" })
  3. Notify: Bash maw hey {MAW or agent-oracle} 'Thread #{id} from {self}: {preview}'
    • If maw hey fails → warn only, don't error (thread already sent)
  4. Confirm: Created channel:{agent} (thread #{id})

Mode 3: One-shot (default)

  1. Compose message from intent
  2. If first arg is #{id} → post directly to that thread ID
  3. Otherwise: arra_threads() → find channel:{agent}, create if missing
  4. Post message to thread
  5. Notify: Bash maw hey {MAW or agent-oracle} 'Thread #{id} from {self}: {preview}'
    • If maw hey fails → warn only, don't error (thread already sent)
  6. arra_thread_read({ threadId }) → show any agent responses
  7. Confirm: Posted to channel:{agent} (thread #{id})
Show full SKILL.md (376 more words)Show less

Mode 4: loop (autonomous conversation)

Like Ralph loop — AI drives the conversation autonomously. No user prompts between turns.

  1. Find or create thread (channel:{agent}, or --new to skip lookup)
  2. Compose opening message from user's intent and post it
  3. Autonomous loop (max 10 iterations): a. arra_thread_read({ threadId }) — check for new messages b. If agent responded: read their response, compose a thoughtful follow-up, post it c. If no new response: compose a follow-up question or probe deeper, post it d. After each exchange, briefly note what you learned e. Stop when: enough insight gathered, conversation circling, or 10 iterations hit
  4. Notify (once, after opening message): Bash maw hey {MAW or agent-oracle} 'Thread #{id} from {self}: {preview}'
    • If maw hey fails → warn only, don't error
  5. Show summary:
    Conversation with {agent} (thread #{id}) — {n} messages, {iterations} turns
    
    Key insights:
    - [insight 1]
    - [insight 2]
  6. Leave thread open for future use

The goal is insight extraction. You are having a conversation on behalf of the human to learn something useful.

Parsing Rules

  • First arg = agent name (lowercase), #id (thread ref), or --list
  • --new = skip lookup, create fresh
  • loop = autonomous conversation (AI drives, no user prompts)
  • --topic "slug" = use topic:{agent}:{slug} instead of channel:{agent}
  • Everything else = the message/intent

Message Composition

CRITICAL: You are the composer. The user gives intent, you write the message.

  • Compose a clear, natural message from the user's intent
  • Post immediately — do NOT ask the user what to say
  • Do NOT use AskUserQuestion for message content
  • Show what you posted after sending

If the message already reads like a direct message (e.g. "What's your status?"), post as-is.

Auto Notification (maw hey)

After posting to a thread, notify the target agent via maw hey:

maw hey {MAW or agent-oracle} 'Thread #{id} from {self}: {first 60 chars of message}'
  • {MAW} = contacts maw field (e.g. "mawjs-oracle") — preferred over {agent}-oracle
  • {self} = current Oracle's name (e.g. "Mother Oracle")
  • {preview} = first ~60 chars of the posted message
  • Runs once per /talk-to invocation (not per loop iteration)
  • Fail-safe: if maw hey errors, log warning and continue — the thread is the source of truth

Important Notes

  • Agent names are always lowercase
  • Thread titles are the routing key — never modify existing thread titles
  • One channel thread per agent (reuse, don't recreate)
  • #{id} lets users reference any thread directly — no lookup needed
  • All messages attributed with role: "human"

ARGUMENTS: $ARGUMENTS

© Soul-Brews-Studio, 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 src/skills/talk-to of Soul-Brews-Studio/arra-oracle-skills-cli.

Open the folder on GitHubat commit 9feea09

Compare with similar skills

Talk To 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.

Talk To compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Talk To this skillSoul-Brews-Studio/arra-oracle-skills-cli123—~2.3kAutomated safety check: PassMIT
Skillsharerunkids/skillshare2.8k—~2.3kAutomated safety check: PassMIT
PicoClaw Agentsipeed/picoclaw30k—~7.2kAutomated safety check: NotesMIT
Autocontext for Hermesgreyhaven-ai/autocontext1.3k—~2.5kAutomated safety check: PassApache-2.0
Dotagentsgetsentry/sentry-wizard295—~900Automated safety check: PassCustom licence
Sandboxed.sh Library ManagementTh0rgal/sandboxed.sh524—~931Automated safety check: PassNone

Similar skills

  • Skillshare

    runkids/skillshare

    Manage skills, agents, extras, hooks, plugins, and MCP connection settings with the Skillshare CLI.

    2.8k GitHub stars~2.3k tokensUpdated today
    Agent WorkflowsAuto-check passed
  • PicoClaw Agent

    sipeed/picoclaw

    Answers questions about running and changing PicoClaw, from onboarding and model selection to MCP server setup, skill loading and scheduled jobs.

    30k GitHub stars~7.2k tokensUpdated yesterday
    Agent WorkflowsAuto-check: notes
  • Autocontext for Hermes

    greyhaven-ai/autocontext

    Lets a Hermes agent run Autocontext scenarios, inspect Hermes curator state, export reusable knowledge and prepare local MLX or CUDA training data through the autoctx CLI.

    1.3k GitHub stars~2.5k tokensUpdated 4 days ago
    Agent WorkflowsAuto-check passed
  • Dotagents

    getsentry/sentry-wizard

    Official

    Manage agent skill dependencies with dotagents. An agent skill from getsentry/sentry-wizard.

    295 GitHub stars~900 tokensUpdated 2 days ago
    Agent WorkflowsAuto-check passed
  • Reads and updates the Sandboxed.sh Library, a git-backed store of skills, agents, commands, tools, rules and MCP servers, through its library tools.

    524 GitHub stars~931 tokensUpdated today
    Agent WorkflowsAuto-check passed
  • Dotagents

    getsentry/dotagents

    Official

    Manage dotagents dependencies and runtime config. An agent skill from getsentry/dotagents.

    251 GitHub stars~1.8k tokensUpdated today
    Agent WorkflowsAuto-check passed

More from Soul-Brews-Studio/arra-oracle-skills-cli

All 31 skills in this repo
  • Dig

    Soul-Brews-Studio/arra-oracle-skills-cli

    Mine Claude Code sessions — timeline, gaps, repo attribution, session history.

    123 GitHub stars~1.8k tokensUpdated 8 days ago
    Auto-check passed
  • Forward

    Soul-Brews-Studio/arra-oracle-skills-cli

    Hand off the current session to the next one. An agent skill from Soul-Brews-Studio/arra-oracle-skills-cli.

    123 GitHub stars~2.9k tokensUpdated 8 days ago
    Auto-check passed
  • Project

    Soul-Brews-Studio/arra-oracle-skills-cli

    Clone and track external repos. An agent skill from Soul-Brews-Studio/arra-oracle-skills-cli.

    123 GitHub stars~1.4k tokensUpdated 8 days ago
    Auto-check passed
  • Round Table

    Soul-Brews-Studio/arra-oracle-skills-cli

    Convene a Round Table (โต๊ะกลม): 3–8 souls — persona files, Oracles, or real people's DNA — debate ONE stakes-bearing decision over exactly 4 rounds (OPEN, DEEPEN, CONVERGE, VOTE), one subagent per…

    123 GitHub stars~2.1k tokensUpdated 8 days ago
    Auto-check passed
  • Soul

    Soul-Brews-Studio/arra-oracle-skills-cli

    Bring an Oracle into the current repo — its soul, ALL its skills and its expertise — by link, never by copy.

    123 GitHub stars~1.3k tokensUpdated 8 days ago
    Auto-check passed
  • Trace

    Soul-Brews-Studio/arra-oracle-skills-cli

    Find projects, code, and knowledge across git history, repos, docs, and Oracle.

    123 GitHub stars~2.9k tokensUpdated 8 days ago
    Auto-check passed

Categories

Questions about Talk To

What does Talk To do?

Talk to another Oracle agent via contacts + threads. An agent skill from Soul-Brews-Studio/arra-oracle-skills-cli. Talk To is an agent skill from Soul-Brews-Studio/arra-oracle-skills-cli. Talk to another Oracle agent via contacts + threads.

When should I use Talk To?

Talk To fits situations like: user says talk to; wants to communicate with another agent (e.g; oracleNet social feed (use /oraclenet); skill management (use /oracle).

How do I install Talk To in Claude Code?

Run `npx skills add Soul-Brews-Studio/arra-oracle-skills-cli --skill talk-to -a claude-code`. Or copy the skill folder (src/skills/talk-to in Soul-Brews-Studio/arra-oracle-skills-cli) into .claude/skills/talk-to in your project. Claude Code loads it when a task matches its description.

How do I install Talk To in Codex?

Run `npx skills add Soul-Brews-Studio/arra-oracle-skills-cli --skill talk-to -a codex`. Or copy the skill folder (src/skills/talk-to in Soul-Brews-Studio/arra-oracle-skills-cli) into .agents/skills/talk-to in your project. Codex loads it when a task matches its description.

Can I use Talk To 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 Soul-Brews-Studio/arra-oracle-skills-cli --skill talk-to -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/talk-to, .gemini/skills/talk-to, .github/skills/talk-to and .opencode/skills/talk-to in your project.

What does Talk To need to run?

Going by SKILL.md and its folder, Talk To needs the command-line tools its instructions call (jq and bash).

Does Talk To 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 Talk To safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Talk To use?

Talk To 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 Talk To use?

About 2.3k tokens (SKILL.md is roughly 9.2k 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 Talk To?

Skills that share tags, products or a category with Talk To: Skillshare (runkids/skillshare, 2.8k stars), PicoClaw Agent (sipeed/picoclaw, 30k stars), Autocontext for Hermes (greyhaven-ai/autocontext, 1.3k stars) and Dotagents (getsentry/sentry-wizard, 295 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Talk To?

Soul-Brews-Studio (a GitHub organization) maintains it in Soul-Brews-Studio/arra-oracle-skills-cli, which has 123 GitHub stars. The repository holds 31 skills in this directory. The repository was last updated on October 3, 2026.

Source: Soul-Brews-Studio/arra-oracle-skills-cli on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.