Claude Code Agent Development
anthropics/claude-plugins-official
Explains how to write agents for Claude Code plugins: the markdown file with YAML frontmatter, trigger descriptions, model and color settings, and system prompt design.
Convene a roundtable of sub-agents that debate a decision or proposal from opposing perspectives — parallel independent statements, anonymous cross-review, then a chaired verdict reporting consensus…
Install Roundtable
$ npx skills add tt-a1i/matt-skills-with-to-goal --skill roundtable -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install tt-a1i/matt-skills-with-to-goal roundtable --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/tt-a1i/matt-skills-with-to-goal.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/engineering/roundtable .claude/skills/roundtable && rm -rf skills-srcUse ~/.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/
Install the "roundtable" agent skill from https://github.com/tt-a1i/matt-skills-with-to-goal/tree/main/skills/engineering/roundtable into .claude/skills/roundtable/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "roundtable", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/tt-a1i/matt-skills-with-to-goal/tree/main/skills/engineering/roundtableType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add tt-a1i/matt-skills-with-to-goal --skill roundtable -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install tt-a1i/matt-skills-with-to-goal roundtable --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/tt-a1i/matt-skills-with-to-goal.git skills-src && mkdir -p .agents/skills && cp -r skills-src/skills/engineering/roundtable .agents/skills/roundtable && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "roundtable" agent skill from https://github.com/tt-a1i/matt-skills-with-to-goal/tree/main/skills/engineering/roundtable into .agents/skills/roundtable/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "roundtable", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add tt-a1i/matt-skills-with-to-goal --skill roundtable -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install tt-a1i/matt-skills-with-to-goal roundtable --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/tt-a1i/matt-skills-with-to-goal.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/skills/engineering/roundtable .cursor/skills/roundtable && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "roundtable" agent skill from https://github.com/tt-a1i/matt-skills-with-to-goal/tree/main/skills/engineering/roundtable into .cursor/skills/roundtable/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "roundtable", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/tt-a1i/matt-skills-with-to-goal.git --path skills/engineering/roundtable--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add tt-a1i/matt-skills-with-to-goal --skill roundtable -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install tt-a1i/matt-skills-with-to-goal roundtable --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/tt-a1i/matt-skills-with-to-goal.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/skills/engineering/roundtable .gemini/skills/roundtable && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "roundtable" agent skill from https://github.com/tt-a1i/matt-skills-with-to-goal/tree/main/skills/engineering/roundtable into .gemini/skills/roundtable/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "roundtable", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install tt-a1i/matt-skills-with-to-goal roundtableInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add tt-a1i/matt-skills-with-to-goal --skill roundtable -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/tt-a1i/matt-skills-with-to-goal.git skills-src && mkdir -p .github/skills && cp -r skills-src/skills/engineering/roundtable .github/skills/roundtable && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "roundtable" agent skill from https://github.com/tt-a1i/matt-skills-with-to-goal/tree/main/skills/engineering/roundtable into .github/skills/roundtable/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "roundtable", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add tt-a1i/matt-skills-with-to-goal --skill roundtable -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install tt-a1i/matt-skills-with-to-goal roundtable --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/tt-a1i/matt-skills-with-to-goal.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/skills/engineering/roundtable .opencode/skills/roundtable && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "roundtable" agent skill from https://github.com/tt-a1i/matt-skills-with-to-goal/tree/main/skills/engineering/roundtable into .opencode/skills/roundtable/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "roundtable", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
roundtable- GitHub stars
- 183
- Token cost
- ~1.2k tokens
- SKILL.md length
- 630 words
- Files
- 2
- Skills in repo
- 6
- Repo updated
- First seen
- Licence
- MIT
At a glance
Convene a roundtable of sub-agents that debate a decision or proposal from opposing perspectives — parallel independent statements, anonymous cross-review, then a chaired verdict reporting consensus…
- Works in 5 steps: Frame the motion → Seat the table → Round one — independent statements → …
- Tasks that involve Subagents
- SKILL.md covers Process and Cost guardrails
- Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md
What it does
Roundtable is an agent skill from tt-a1i/matt-skills-with-to-goal. Convene a roundtable of sub-agents that debate a decision or proposal from opposing perspectives — parallel independent statements, anonymous cross-review, then a chaired verdict reporting consensus and dissent.
Its SKILL.md is about 1.2k tokens, which your agent loads only when the skill is triggered. The skill folder holds 2 other files (for example `agents/openai.yaml`).
It sits in Agent Workflows, covering Subagents. The repository describes itself as: Planning → verifiable goals → fresh-session implementation for AI coding agents. Based on mattpocock/skills v1.1. The licence is MIT.
When your agent uses it
- Tasks that involve Subagents
Example prompts
- “/roundtable”
Workflow steps
5 steps, taken from the step headings in SKILL.md.
What it can do on your machine
Read from SKILL.md and the folder at commit 7b44372. 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
Roundtable loads about 1.2k tokens when it runs. Until then it costs about 56 tokens; SKILL.md has 630 words of instructions outside code blocks.
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 tt-a1i/matt-skills-with-to-goal at commit 7b44372, republished under its MIT licence (© tt-a1i). 630 words, ~1,176 tokens.
Download SKILL.mdSave it as
.claude/skills/roundtable/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.- name
- roundtable
- description
- Convene a roundtable of sub-agents that debate a decision or proposal from opposing perspectives — parallel independent statements, anonymous cross-review, then a chaired verdict reporting consensus and dissent.
- disable-model-invocation
- true
Pressure-test a decision, plan, or proposal by dispatching parallel sub-agents that argue it from deliberately opposed perspectives, then synthesising their debate into one verdict that preserves dissent.
Scope: decisions and designs only. Reviewing a diff is /code-review's job — don't point the roundtable at code changes.
Process
1. Frame the motion
Restate the user's question as a single debatable motion — a claim the table can support or oppose ("We should extract X into its own service"), not an open question ("What should we do about X?"). If the user's ask is too vague to frame, ask them to sharpen it first; if it's still fog, suggest running /grilling before convening the table.
Then assemble the evidence pack yourself, in the main thread: relevant file paths and excerpts, constraints, decisions already made, anything the user attached. Finding facts is the chair's job, not the seats' — sub-agents don't hunt on their own, so the whole table debates the same evidence.
Before dispatching anyone, confirm the motion is one sentence and the evidence pack is complete enough to argue from. A muddled motion should fail here — not inside four parallel sub-agents.
2. Seat the table
Default four seats:
- Skeptic — assumes the motion fails and argues why; enforces YAGNI; hunts hidden assumptions.
- Architect — argues from long-term structure: module depth, coupling, cost of evolution.
- User Advocate — argues from the experience of whoever uses or calls the result; real workflows over hypothetical ones.
- Pragmatist — argues from the shortest viable path: maintenance burden, opportunity cost, what ships this week.
The user may recompose the table for the topic ("seat a table of security experts") — keep it to 3–5 seats. Every seat brief, default or custom, must carry this line verbatim: "Your job is to make the strongest case from this perspective, even if your all-things-considered judgement would land elsewhere." That line is the first defence against convergence.
3. Round one — independent statements
Dispatch N sub-agents in parallel, one per seat. Each gets an identical input pack except for its seat brief — include:
- The motion, verbatim.
- The full evidence pack — the sub-agent has no other access to it.
- Its seat brief, including the strongest-case line from step 2.
- The output contract: "State your position (support / oppose / support-with-conditions), your three strongest arguments, and the single biggest risk you see. Under 300 words."
Isolation is the point: no seat sees another seat's statement, so no one anchors on the first opinion voiced.
Show full SKILL.md (229 more words)Show less
4. Round two — anonymous cross-review
Skip this round only when the user asked for a quick roundtable.
Strip the seat names from the round-one statements, shuffle their order, and label them Position A, B, C, D. Dispatch N sub-agents in parallel again — each gets:
- The motion and evidence pack, as before.
- Its own seat brief, plus which anonymized position was its own.
- All anonymized positions in full.
- The brief: "Rebut the position you most disagree with, quoting the claim you're attacking. Concede any point in your own position that another position genuinely weakened. Rank A–D by persuasiveness, best to worst. Under 300 words."
Anonymization makes seats attack arguments, not personas — a rebuttal aimed at "the Skeptic" is theatre; one aimed at an unattributed claim has to engage the claim.
5. Chair's synthesis
The main thread de-anonymizes and judges. Judging is not averaging: a position isn't wrong because it's outnumbered, and dissent that no rebuttal actually defeated goes in the report intact, attributed to its seat. Fixed output shape:
## Verdict — the recommendation, one paragraph
## Consensus — points every seat agreed on
## Dissent worth weighing — undefeated objections, attributed to their seats
## Vote table — seat × position × cross-review ranking
## What would change the verdict — facts that would flip itDon't act on the verdict — the roundtable advises; the user decides.
Cost guardrails
A full roundtable is 2N sub-agent runs (N statements + N cross-reviews); quick mode is N. Two rounds is the ceiling — no open-ended debate loops. If the table hasn't earned a verdict after cross-review, that itself is the finding: report the unresolved split and what evidence would resolve it.
© tt-a1i, 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 skills/engineering/roundtable of tt-a1i/matt-skills-with-to-goal.
- SKILL.md
- agents/openai.yaml
Open the folder on GitHubat commit 7b44372
Compare with similar skills
Roundtable 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.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Roundtable this skilltt-a1i/matt-skills-with-to-goal | 183 | — | ~1.2k | Automated safety check: Pass | MIT | |
| Claude Code Agent Developmentanthropics/claude-plugins-official | 38k | 8 repos | ~2.8k | Automated safety check: Pass | Apache-2.0 | |
| Subagent Driven DevelopmentAsvarox/allkaraoke | 261 | 38 repos | ~1.2k | Automated safety check: Pass | None | |
| Dispatching Parallel Agentsultralisp/ultralisp | 258 | 41 repos | ~1.5k | Automated safety check: Pass | None | |
| Paseo Advisor Second Opiniongetpaseo/paseo | 20k | 1 repos | ~756 | Automated safety check: Pass | Custom licence | |
| Task Observerrebelytics/one-skill-to-rule-them-all | 3.2k | 1 repos | ~12k | Automated safety check: Pass | CC-BY-4.0 |
Similar skills
- Official
Subagent Driven Development
Asvarox/allkaraoke
A skill your agent uses when executing implementation plans with independent tasks in the current session
Dispatching Parallel Agents
ultralisp/ultralisp
A skill your agent uses when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies
Paseo Advisor Second Opinion
getpaseo/paseo
Launches one separate agent through Paseo to give a second opinion on the current task, with a self-contained briefing and no permission to edit files.
Task Observer
rebelytics/one-skill-to-rule-them-all
Monitors task execution for skill improvement opportunities.
O2 Review Loop
openobserve/openobserve
Splits a change into planner, coder and independent reviewer roles: you confirm a spec, a subagent implements it, and a separate reviewer checks each round's local WIP commit.
More from tt-a1i/matt-skills-with-to-goal
Goal Crafter
tt-a1i/matt-skills-with-to-goal
Craft verifiable, tight goal prompts for AI coding agents (Claude Code /goal, Codex Automations, Pi), either from a standalone task interview or from an already-approved to-goal handoff.
Spec Executor
tt-a1i/matt-skills-with-to-goal
Execute approved, bounded work in an isolated implementation conversation and return a structured evidence receipt.
To Goal
tt-a1i/matt-skills-with-to-goal
Turn approved planning evidence or partially completed work into a portable, verifiable execution goal.
Execute Spec In Fork
tt-a1i/matt-skills-with-to-goal
Orchestrate approved, bounded work through a same-directory Codex App fork: create and name the execution task, ask it to run spec-executor, route decisions through Codex Task Messenger, validate…
Handoff
tt-a1i/matt-skills-with-to-goal
Compact the current conversation into a portable handoff document, with an optional live clarification path for compatible Codex tasks.
Related
Categories
Questions about Roundtable
What does Roundtable do?
Convene a roundtable of sub-agents that debate a decision or proposal from opposing perspectives — parallel independent statements, anonymous cross-review, then a chaired verdict reporting consensus…. Roundtable is an agent skill from tt-a1i/matt-skills-with-to-goal. Convene a roundtable of sub-agents that debate a decision or proposal from opposing perspectives — parallel independent statements, anonymous cross-review, then a chaired verdict reporting consensus and dissent.
When should I use Roundtable?
Roundtable fits situations like: tasks that involve Subagents.
How do I install Roundtable in Claude Code?
Run `npx skills add tt-a1i/matt-skills-with-to-goal --skill roundtable -a claude-code`. Or copy the skill folder (skills/engineering/roundtable in tt-a1i/matt-skills-with-to-goal) into .claude/skills/roundtable in your project. Claude Code loads it when a task matches its description.
How do I install Roundtable in Codex?
Run `npx skills add tt-a1i/matt-skills-with-to-goal --skill roundtable -a codex`. Or copy the skill folder (skills/engineering/roundtable in tt-a1i/matt-skills-with-to-goal) into .agents/skills/roundtable in your project. Codex loads it when a task matches its description.
Can I use Roundtable 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 tt-a1i/matt-skills-with-to-goal --skill roundtable -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/roundtable, .gemini/skills/roundtable, .github/skills/roundtable and .opencode/skills/roundtable in your project.
What does Roundtable need to run?
SKILL.md names no scripts, command-line tools or credentials: Roundtable is instructions for the agent only.
Does Roundtable 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 Roundtable 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 Roundtable use?
Roundtable 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 Roundtable use?
About 1.2k tokens (SKILL.md is roughly 4.7k 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 Roundtable?
Skills that share tags, products or a category with Roundtable: Claude Code Agent Development (anthropics/claude-plugins-official, 38k stars), Subagent Driven Development (Asvarox/allkaraoke, 261 stars), Dispatching Parallel Agents (ultralisp/ultralisp, 258 stars) and Paseo Advisor Second Opinion (getpaseo/paseo, 20k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Roundtable?
tt-a1i (a GitHub user) maintains it in tt-a1i/matt-skills-with-to-goal, which has 183 GitHub stars. The repository holds 6 skills in this directory. The repository was last updated on September 14, 2026.
Source: tt-a1i/matt-skills-with-to-goal on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.