Agent skill

BTW Side-Conversation Workflow

by dbachelder in dbachelder/pi-btw

Runs a side question or brainstorm in a parallel thread so it does not derail the main agent's current work, then optionally hands the result back.

MITAuto-check passedAgent Workflows

Install BTW Side-Conversation Workflow

skills CLI
$ npx skills add dbachelder/pi-btw --skill btw -a claude-code

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

GitHub CLI
$ gh skill install dbachelder/pi-btw btw --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/dbachelder/pi-btw.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/btw .claude/skills/btw && 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
btw
GitHub stars
207
Token cost
~1.2k tokens
SKILL.md length
546 words
Files
1
Skills in repo
1
Repo updated
First seen
Licence
MIT

At a glance

Runs a side question or brainstorm in a parallel thread so it does not derail the main agent's current work, then optionally hands the result back.

  • Asking a side question while the main agent keeps working on its task
  • SKILL.md covers When to use BTW, Commands, How to guide the user and Recommendation rules, plus 2 more sections
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md
  • Brainstorming an alternative approach without derailing the current turn

What it does

Offers several side-thread commands for different needs: a plain `/btw <question>` for a quick aside, `/btw --save <question>` when the exchange should appear as a visible note in the session transcript, `/btw:new` to start a fresh thread that still inherits the current session's context, and `/btw:tangent` for a side thread with no inherited context at all. `/side` is kept as an alias of `/btw` for people coming from other tools, sharing the same thread and settings.

A `/btw:ask` variant runs a side question that inherits context but is restricted to read-only tools (`read`, `grep`, `find`, `ls`), with no ability to edit or run shell commands. Once a side thread has run its course, `/btw:inject` hands the full discussion back to the main agent to act on, while `/btw:summarize` hands back a condensed version instead.

When your agent uses it

  • Asking a side question while the main agent keeps working on its task
  • Brainstorming an alternative approach without derailing the current turn
  • Running a read-only side investigation that cannot change any files
  • Handing a side thread's conclusion back into the main conversation

Example prompts

  • “/btw would a queue-based approach work better here?”
  • “/btw:ask look through the auth module and tell me how sessions expire.”
  • “/btw:summarize send the conclusion back to the main thread.”

What it can do on your machine

Read from SKILL.md and the folder at commit d0d1ba5. 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

    No scripts in the folder and no shell commands in SKILL.md.

    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

BTW Side-Conversation Workflow loads about 1.2k tokens when it runs. Until then it costs about 53 tokens; SKILL.md has 546 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~53
When it runs · the whole SKILL.md, loaded when a task matches
~1.2k

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 dbachelder/pi-btw at commit d0d1ba5, republished under its MIT licence (© dbachelder). 546 words, ~1,237 tokens.

Download SKILL.mdSave it as .claude/skills/btw/SKILL.md (or your agent's skills folder).
name
btw
description
Helps you use the /btw side-conversation workflow effectively. Use when you want to think in parallel, ask side questions without interrupting ongoing work, or inject a side thread back into the main agent.

BTW

Use this skill when the user wants to work in parallel with the main agent instead of derailing the current turn.

When to use BTW

Prefer the BTW workflow when the user wants to:

  • ask a side question while the main agent keeps working
  • brainstorm or compare options without interrupting the current run
  • prepare a plan or summary before handing it back to the main agent
  • keep exploratory discussion out of the main transcript/context

Commands

Use these commands in your guidance to the user:

text
/btw <question>
/side <question>
/btw --save <question>
/btw:new [question]
/btw:tangent <question>
/btw:tangent --save <question>
/btw:ask <question>
/btw:ask --save <question>
/btw:clear
/btw:model [<provider> <model> <api> | clear]
/btw:thinking [<level> | clear]
/btw:inject [instructions]
/btw:summarize [instructions]

/side is an alias for /btw and shares the same thread, overlay, and settings. /btw is canonical; the lifecycle commands stay in the /btw:* namespace.

How to guide the user

For a quick side question

Recommend:

text
/btw <question>

Use this when the user wants an immediate aside and does not need a visible saved note.

Users coming from Codex may reach for /side instead; it is an alias for /btw and can be used anywhere /btw is used.

For a saved one-off note

Recommend:

text
/btw --save <question>

Use this when the user wants the exchange to appear as a visible BTW note in the session transcript.

For a fresh side thread

Recommend:

text
/btw:new

or

text
/btw:new <question>

Use this when the previous BTW discussion is no longer relevant, but you still want the new side thread to inherit the current main-session context.

For a contextless tangent thread

Recommend:

text
/btw:tangent <question>

or

text
/btw:tangent --save <question>

Use this when the user wants a side conversation that does not include the current main-session context.

For an enforced read-only side question

Recommend:

text
/btw:ask <question>

or

text
/btw:ask --save <question>

Use this when the user wants a side question that inherits the current main-session context but must not be able to change anything. The read-only thread only has read, grep, find, and ls; it never has bash, edit, or write.

To hand the full thread back to the main agent

Recommend:

text
/btw:inject <instructions>

Use this when the exact discussion matters and the user wants the main agent to act on it.

Show full SKILL.md (227 more words)Show less
To hand back a condensed version

Recommend:

text
/btw:summarize <instructions>

Use this when the thread is long and only the distilled outcome should go back into the main agent.

To make BTW cheaper or faster than the main thread

Recommend:

text
/btw:model <provider> <model> <api>
/btw:thinking <level>

Use these when the main thread should keep its current model or thinking level, but BTW should run with a different cost/speed profile.

Recommendation rules

  • Prefer /btw over normal chat when the user explicitly wants a side conversation.
  • Prefer /btw:tangent when the user wants that side conversation to be contextless.
  • Prefer /btw:ask when the user wants a side conversation that cannot modify the workspace.
  • Prefer /btw:summarize over /btw:inject for long exploratory threads.
  • Prefer /btw:inject when precise wording, detailed tradeoffs, or a full plan matters.
  • Suggest /btw:new before starting a totally unrelated side topic when main-session context is still useful.
  • Suggest /btw:clear when the widget/thread should be dismissed.
  • Suggest /btw:model or /btw:thinking when the user wants BTW to be cheaper, faster, or less deliberative than the main thread.

Response style

When helping the user use BTW:

  • give the exact slash command to run
  • explain briefly why that command fits
  • keep the guidance short and operational

Examples

Example: brainstorm while coding continues
text
/btw what are the risks of switching this to optimistic updates?
Example: create a clean new thread
text
/btw:new sketch a safer migration plan
Example: start a contextless tangent
text
/btw:tangent think through this from first principles without using the current chat context
Example: ask a read-only side question
text
/btw:ask what does the token refresh path do without changing anything?
Example: send the result back
text
/btw:summarize implement the recommended migration plan
Example: make BTW cheaper than the main thread
text
/btw:model openai gpt-5-mini openai-responses
/btw:thinking low

© dbachelder, 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 skills/btw of dbachelder/pi-btw.

Open the folder on GitHubat commit d0d1ba5

Compare with similar skills

BTW Side-Conversation Workflow 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.

BTW Side-Conversation Workflow compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
BTW Side-Conversation Workflow this skilldbachelder/pi-btw207—~1.2kAutomated safety check: PassMIT
Orca CLIstablyai/orca87k2 repos~593Automated safety check: PassMIT
Paseo Agent Handoffgetpaseo/paseo20k1 repos~606Automated safety check: PassCustom licence
MemPalace Task HandoffMemPalace/mempalace59k—~1.9kAutomated safety check: PassMIT
Agent Manager Fleet TUIYoanWai/agent-manager572—~1kAutomated safety check: PassApache-2.0
Seat Continuity and Handovermvschwarz/openrig5.5k—~2.1kAutomated safety check: PassApache-2.0

Similar skills

  • Orca CLI

    stablyai/orca

    Operate Orca-managed worktrees, folder contexts, terminals, repos, automations, artifacts, skill sharing, worktree comments, and Orca's embedded browser…

    87k GitHub starsUsed in 2 repos~593 tokens
    Agent WorkflowsAuto-check passed
  • Paseo Agent Handoff

    getpaseo/paseo

    Hands off the current task, including context, decisions and failed attempts, to a fresh agent through Paseo by writing a self-contained briefing prompt and launching that agent.

    20k GitHub starsUsed in 1 repo~606 tokens
    Agent WorkflowsAuto-check passed
  • MemPalace Task Handoff

    MemPalace/mempalace

    Creates, hands off, claims, executes and closes agent tasks through the MemPalace logstream, with approval of the exact task before it is recorded.

    59k GitHub stars~1.9k tokensUpdated today
    Agent WorkflowsAuto-check passed
  • Agent Manager Fleet TUI

    YoanWai/agent-manager

    Runs several coding-agent CLIs as real tmux sessions in one terminal UI, color-coded by whether each is working, waiting, idle or blocked.

    572 GitHub stars~1k tokensUpdated today
    Agent WorkflowsAuto-check passed
  • Separates a stable agent seat's identity from its changing occupant, and records honest, two-part provenance whenever one occupant replaces another.

    5.5k GitHub stars~2.1k tokensUpdated today
    Agent WorkflowsAuto-check passed
  • Harness Engineering

    10xChengTu/harness-engineering

    Set up and improve harness engineering (AGENTS.md, docs/, lint rules, eval systems, project-level prompt engineering) for AI-agent-friendly codebases.

    102 GitHub starsUsed in 1 repo~1k tokens
    Agent WorkflowsAuto-check passed

Categories

Questions about BTW Side-Conversation Workflow

What does BTW Side-Conversation Workflow do?

Runs a side question or brainstorm in a parallel thread so it does not derail the main agent's current work, then optionally hands the result back. Offers several side-thread commands for different needs: a plain `/btw <question>` for a quick aside, `/btw --save <question>` when the exchange should appear as a visible note in the session transcript, `/btw:new` to start a fresh thread that still inherits the current session's context, and `/btw:tangent` for a side thread with no inherited context at all. `/side` is kept as an alias of `/btw` for people coming from other tools, sharing the same thread and settings.

When should I use BTW Side-Conversation Workflow?

BTW Side-Conversation Workflow fits situations like: asking a side question while the main agent keeps working on its task; brainstorming an alternative approach without derailing the current turn; running a read-only side investigation that cannot change any files; handing a side thread's conclusion back into the main conversation.

How do I install BTW Side-Conversation Workflow in Claude Code?

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

How do I install BTW Side-Conversation Workflow in Codex?

Run `npx skills add dbachelder/pi-btw --skill btw -a codex`. Or copy the skill folder (skills/btw in dbachelder/pi-btw) into .agents/skills/btw in your project. Codex loads it when a task matches its description.

Can I use BTW Side-Conversation Workflow 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 dbachelder/pi-btw --skill btw -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/btw, .gemini/skills/btw, .github/skills/btw and .opencode/skills/btw in your project.

What does BTW Side-Conversation Workflow need to run?

SKILL.md names no scripts, command-line tools or credentials: BTW Side-Conversation Workflow is instructions for the agent only.

Does BTW Side-Conversation Workflow 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 BTW Side-Conversation Workflow 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 BTW Side-Conversation Workflow use?

BTW Side-Conversation Workflow 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 BTW Side-Conversation Workflow use?

About 1.2k tokens (SKILL.md is roughly 4.9k 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 BTW Side-Conversation Workflow?

Skills that share tags, products or a category with BTW Side-Conversation Workflow: Orca CLI (stablyai/orca, 87k stars), Paseo Agent Handoff (getpaseo/paseo, 20k stars), MemPalace Task Handoff (MemPalace/mempalace, 59k stars) and Agent Manager Fleet TUI (YoanWai/agent-manager, 572 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains BTW Side-Conversation Workflow?

dbachelder (a GitHub user) maintains it in dbachelder/pi-btw, which has 207 GitHub stars. The repository was last updated on October 1, 2026.

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