Claude-Pipeline: Agent for Claude Code

.claude/agents/adversarial-coordinator.md

adversarial-coordinator is an agent for Claude Code from TheAstrelo/Claude-Pipeline. It costs 28 tokens per session (1,131 once invoked), scanned A, original, MIT.

A design-review agent that examines a proposed technical design from three different viewpoints before coding starts.

In plain words
What is it for?
Use it to review the decisions in a design document and write a combined critique that separates shared concerns from issues raised by only one reviewer.
Why use it?
It helps reveal hidden assumptions, architectural problems, scaling concerns, and implementation risks early, when they are cheaper to fix.

Agent for Claude Code

Written for Claude Code: installed under .claude/. Also seen: model in frontmatter.

This is TheAstrelo/Claude-Pipeline's own configuration. It tells Claude Code how to work on Claude-Pipeline itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything Claude-Pipeline configures →

Reuse

Borrowing it

Nothing to install: this file belongs to TheAstrelo/Claude-Pipeline. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/TheAstrelo/Claude-Pipeline/main/.claude/agents/adversarial-coordinator.md
Clone the repo
git clone --depth 1 https://github.com/TheAstrelo/Claude-Pipeline

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for adversarial-coordinator

README.md
[![agentmods](https://agentmods.dev/badge/agents/theastrelo/claude-pipeline/adversarial-coordinator/github.svg)](https://agentmods.dev/agents/theastrelo/claude-pipeline/adversarial-coordinator)
Your own site
<a href="https://agentmods.dev/agents/theastrelo/claude-pipeline/adversarial-coordinator"><img src="https://agentmods.dev/badge/agents/theastrelo/claude-pipeline/adversarial-coordinator/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for adversarial-coordinator

Your own site · 80×15
<a href="https://agentmods.dev/agents/theastrelo/claude-pipeline/adversarial-coordinator"><img src="https://agentmods.dev/badge/agents/theastrelo/claude-pipeline/adversarial-coordinator.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 28 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,131 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00028 $0.01131
Opus 5 $0.00014 $0.00566
Sonnet 5 $0.00006 $0.00226
Haiku 4.5 $0.00003 $0.00113

Measured 9d ago against content hash 43a9a8672719, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

adversarial-coordinator scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/agents/adversarial-coordinator.md · 153 lines

How it starts

The opening of the file, as written. The whole thing — 153 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are the Adversarial Coordinator agent for the RDO project — a B2B go-to-market intelligence platform built with Next.js, TypeScript, MUI, and PostgreSQL.

Your Job

Stress-test a technical design by critiquing it from multiple perspectives. Surface blind spots, untested assumptions, and implementation risks before any code is written.

Process

  1. Read the design — Load .pipeline/artifacts/current/design.md.
  2. Extract key decisions — List every architectural/technical decision made.
  3. Run 3 critique passes — Each with a different critical perspective.
  4. Synthesize findings — Identify consensus issues vs. single-perspective concerns.
  5. Output critique — Write to .pipeline/artifacts/current/critique.md.

The Three Critics

Critic 1: The Architect

Perspective: "What architectural flaws exist?"

Focus on:

  • Scalability: Does this approach scale with data/users?
  • Coupling: Are components too tightly coupled?
  • Consistency: Does this match existing patterns in the codebase?
  • Extensibility: Will this be painful to extend later?
  • Performance: Are there N+1 queries, missing indexes, or inefficient patterns?

Questions to ask:

  • "What happens when this table has 1M rows?"
  • "How would we add [reasonable extension] later?"
  • "Does this introduce circular dependencies?"

Critic 2: The Skeptic

Perspective: "What assumptions are untested?"

Focus on:

  • Edge cases: Empty data, null values, concurrent access
  • Error states: Network failures, API rate limits, timeouts
  • User behavior: Rapid clicks, back button, multiple tabs
  • Data integrity: Race conditions, partial failures
  • Security: Auth bypass, injection, privilege escalation

Questions to ask:

  • "What if the external API is down?"
  • "What if two users modify the same record simultaneously?"
  • "What if the user has no data for this feature?"

Critic 3: The Implementer

Perspective: "What's ambiguous or will cause bugs?"

Focus on:

  • Clarity: Are interfaces fully specified?
  • Types: Are TypeScript types complete or too loose?
  • State: Is state management clear?
  • Testing: How will this be tested?
  • Dependencies: Are all imports/packages identified?

Read the full file on GitHub · 153 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 9d ago First seen · 153 lines · 28 tokens per session scan A 43a9a8672719

Subscribe to this mod's changes

adversarial-coordinator is an agent published in the GitHub repository TheAstrelo/Claude-Pipeline (44 stars, last pushed 4d ago), licensed MIT. It adds 28 tokens to every session and 1,131 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other agents, from other repositories

spec-fidelity-auditor

Conformance of a change to its own specification (ticket, acceptance criteria, spec KB, solution plan, resolved decisions): a per-change requirements-traceability matrix swept both ways. Forward: every criterion, exact string, event name, numeric value, and named exclusion delivered, with evidence at the promised…

waqas1412/claude-harness · 186 tokens

data-flow-timing-auditor

Cross-file data-flow timing and staleness: signals read before they settle, proxy gates (timers or render flags standing in for data-readiness), one-shot consumers (analytics, seeds, redirects, caches, queue acks) snapshotting eventually-consistent state, init/hydration order, missed-event races, non-convergent…

waqas1412/claude-harness · 145 tokens

design-principles-advisor

Structural design principles: SOLID, GRASP, CUPID, coupling/cohesion, and SoC/composition-over-inheritance/LoD/CoC/CQS/POLA. Two gates: PLAN (shape a sound design) and VERIFY (audit a design or diff). Read-only; ranked violations and strengths with fixes, plus an anti-over-engineering guard. Not…

waqas1412/claude-harness · 102 tokens

developer-reviewer

Correctness and test review of a diff: logic bugs, nil/empty/collection-ordering/boundary/timezone, invariants and contracts, test coverage (red to green), AGENTS.md / CLAUDE.md compliance. Two gates: PLAN (risk and test plan) and VERIFY (adversarial diff review). Read-only; returns findings with severity and a fix.…

waqas1412/claude-harness · 199 tokens

compatibility-auditor

Whether a change breaks an existing consumer, and whether the PR's breaking-change assertion is true. Sweeps outward from the diff to every consumer of what it touched: API request/response and status contracts, published schemas (OpenAPI, TypeSpec, GraphQL), database schema and migration reversibility, shared…

waqas1412/claude-harness · 151 tokens

security-auditor

Security review of a diff along the axes a static scanner cannot judge: authorization and tenant/customer scoping, authentication and token/session handling, secrets and credential exposure, sensitive data reaching logs, analytics, URLs or client bundles, injection and deserialization at real trust boundaries, and…

waqas1412/claude-harness · 146 tokens