cd /news/ai-agents/show-hn-pantheon-ai-vs-ai-one-writes… · home topics ai-agents article
[ARTICLE · art-28084] src=github.com ↗ pub= topic=ai-agents verified=true sentiment=↑ positive

Show HN: Pantheon – AI vs AI: one writes the code, the other attacks it

A developer released Pantheon, an open-source multi-agent harness for Claude Code that runs coding tasks through parallel implementations and adversarial verification to catch bugs that single-pass models miss. The tool uses a pipeline of planning, parallel implementation, adversarial testing by a second AI, and synthesis, and is available in two versions: pantheon (Claude-only) and pantheon-x (cross-model with GPT-5.5).

read6 min publishedJun 15, 2026

Two Claude Code skills that run a hard coding task through a multi-agent harness instead of a single model pass: plan → N parallel implementations → adversarial verification → judge. The point isn't a smarter model — it's that a second (and third) implementation, plus an independent reviewer whose job is to break the result, catches bugs a single pass ships green.

It's a packaging of well-worn techniques — best-of-N sampling, tool-integrated self-correction, and LLM-as-judge / adversarial verification — wired into one /pantheon

command so you don't reassemble them by hand each time. This is scaffolding around the model, not a change to it: it won't rescue a task the model fundamentally can't reason about, but it reliably tightens correctness on coding work whose answer you can express as tests.

The harness runs a deterministic pipeline:

Plan ──▶ Implement (×N parallel) ──▶ Verify (adversarial ×V) ──▶ Synthesize
 │            │ each self-corrects            │ try to BREAK each      │ judge picks winner
 1 planner    │ against its own tests (T1)    │ green build            │ + grafts best ideas
              N builders                       reviewers

Plan— derive a tight spec, a test plan thatdefinescorrectness, and N distinct strategies (before any code).Implement— N builders implement different strategies in parallel; each runs its own tests and self-corrects on failure (tool-integrated self-verification, up to 5 iterations).Verify— independent adversarial reviewers try tobreakeach green build; a build refuted by a majority is dropped.Synthesize— a judge picks the winner and lists superior ideas worth grafting from the runners-up.

The value: a build can pass its own tests yet still be wrong. The adversarial layer catches defects the self-written tests miss, instead of rubber-stamping a green build.

Skill Adversarial verifier Requirements
pantheon
Claude itself (independent agents) Paid Claude Code plan + Workflows (see below)
pantheon-x
GPT-5.5 via Codex plugin (cross-model)
Above + OpenAI Codex plugin (codex:codex-rescue )

pantheon-x

is the stronger setting: the implementation written by Claude is attacked by a different model, which shrinks single-model blind spots (the same mistake slipping past a same-model verifier). If you don't have Codex/GPT-5.5, use pantheon

.

Both skills share the same harness (pantheon-class.js

); they differ only in the crossModelVerify

flag.

These skills drive Claude Code's Workflow orchestration engine, so a stock/Free setup is not enough:

Claude Code ≥ v2.1.154 on apaid plan— Pro, Max, Team, or Enterprise (also Bedrock / Vertex / Foundry).** Not available on the Free tier.**- On Pro, enable it once:/config

→ turn onDynamic workflows. the cross-model verifier runs as thepantheon-x

only:codex:codex-rescue

subagent, which ships in OpenAI'sCodex pluginnotstock Claude Code. A logged-incodex

CLI alone doesnot register it. Install the plugin:plus a ChatGPT subscription (or

/plugin marketplace add openai/codex-plugin-cc
/plugin install codex@openai-codex

OPENAI_API_KEY

) and thecodex

CLI on PATH.Ifcodex:codex-rescue

isn't installed, usepantheon

insteadpantheon-x

would otherwise silently skip the adversarial pass and pass every build.

Skills and subagents themselves are stock Claude Code features; no extra setup beyond the above.

Clone into your Claude Code skills directory (personal install):

git clone https://github.com/lolu1032/pantheon-skills.git
cp -R pantheon-skills/pantheon       ~/.claude/skills/pantheon
cp -R pantheon-skills/pantheon-x     ~/.claude/skills/pantheon-x

Or for a single project, copy into <project>/.claude/skills/

.

In Claude Code:

/pantheon    <a hard implementation task whose correctness is testable>
/pantheon-x  <same, but GPT-5.5 does the adversarial verification>

Example:

/pantheon Add idempotency-key handling to the payments module so concurrent requests can't double-charge. Tests: pnpm test (vitest)

Claude collects the parameters (task

, workdir

, lang

  • test command, variants

, verifiers

) and launches the harness as a background Workflow, then reports: per-variant test results, which builds the adversarial pass broke, and the final winner with its rationale and grafting suggestions.

arg default notes
task
one-paragraph requirement + acceptance criteria (expressible as tests)
workdir
/tmp/pantheon-<name>
absolute path; a real repo or a scratch dir
lang
Python/unittest language + the exact test command for your stack
variants
3 bump to 5 for harder problems
verifiers
2 bump to 3 to be stricter (majority refutation drops a build)
crossModelVerify
false (pantheon ) / true (pantheon-x )
route adversarial verify to GPT-5.5/Codex

Not a daemon. Each invocation runs once to completion and exits — zero cost when idle.- A run spends real tokens. A representative run is ~11 subagents and a few hundred K to ~1M tokens end-to-end, ~6–10 min wall-clock; heavier settings ( variants=5

,verifiers=3

, cross-model) cost more. On Pro/Max it draws from your usage quota; on metered API access, budget a few dollars per run and up.Route only the hardest 10–20% of tasks here— use plain Opus for the rest. - This buys correctness on testable work, not raw model intelligence. If a task isn't expressible as tests, the adversarial layer has little to grip and the overhead isn't worth it. - Coding/agentic productivity only. Not a tool for bypassing safety gates (cybersecurity/biology capability restrictions).

Isn't this just a prompt wrapper? There's no model change — it's orchestration, yes. The non-trivial part is the adversarial step: an independent agent (a different model in pantheon-x

) whose job is to break a build rather than confirm it. That's what catches defects the builder's own green tests rubber-stamp. The value is the harness shape, not a secret prompt.

Do you have benchmarks vs. plain Opus? No formal benchmark yet — treat the description as mechanism, not a measured delta. The value is in the adversarial step: a build can pass its own tests and still be wrong, and an independent reviewer catches what the self-written tests rubber-stamp. If you run a head-to-head, I'd genuinely like to see the numbers.

What does a run cost? A few hundred K to ~1M tokens and ~6–10 min at default settings; more for variants=5

/ verifiers=3

/ cross-model. It's meant for the hardest 10–20% of tasks, not everyday edits. See Cost & scope.

It says "Workflow tool not found" / nothing happens. You're likely on the Free tier, or haven't enabled workflows. See Requirements — needs a paid plan and, on Pro, /config

Dynamic workflows.

Why route verification to GPT-5.5 / another vendor's model? Same-model verifiers share blind spots — a mistake the builder makes, a same-model reviewer tends to miss too. A different model is a cheap way to break that correlation. It's optional: pantheon

runs Claude-on-Claude and still helps.

Solo project, as-is, best-effort. Issues and PRs are welcome, but maintenance comes with no guarantees or SLA — I may not get to everything. It's MIT-licensed, so forking is a first-class option if you want to take it further.

── more in #ai-agents 4 stories · sorted by recency
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/show-hn-pantheon-ai-…] indexed:0 read:6min 2026-06-15 ·