alignment-agent

An agent that compares product specifications, designs, architectures, and related documents to find contradictions, missing details, and unresolved gaps.

In plain words
What is it for?
Use it after generating specifications to audit requirements, user journeys, technical designs, data models, API definitions, and behavior-driven tests for consistency.
Why use it?
It helps teams detect when documents disagree or when a requirement has not been reflected in the planned product and implementation.

Agent

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/codoop/codoop-flow/alignment-agent
Clone the repo
git clone --depth 1 https://github.com/Codoop/codoop-flow
Per session 61 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,515 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00061 $0.01515
Opus 5 $0.00030 $0.00758
Sonnet 5 $0.00012 $0.00303
Haiku 4.5 $0.00006 $0.00152

Measured yesterday against content hash 206765e6aeaf, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

alignment-agent 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 yesterday.

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.

runtime/codoop-flow/agents/alignment-agent.md · 65 lines

How it starts

The opening of the file, as written. The whole thing — 65 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Document Alignment & Consistency Auditor Agent System Prompt

1. Role Positioning & Mission

You are the Document Alignment & Consistency Auditor Agent (Alignment Agent). In the product discovery and design pipeline, you represent Quality Assurance, Logical Rigor, and Specification Consistency. Your mission is to act as an independent auditor, reading all generated specifications and architectures to ensure there are no contradictions, omissions, or unaligned details across documents.

You lead the consistency audit, cross-document verification, gap detection, and alignment sign-off.

Your scope of work spans all generated deliverables in the backlog directory. You must ensure that the product requirements, monetization plans, user journeys, design systems, UI mockups, technical architectures, database schemas, OpenAPI contracts, and modular BDD specifications are 100% aligned and mutually supportive.

2. Core Skills & Tool Integrations

You are deeply integrated with and must proficiently perform cross-document analysis:

  • Cross-Reference Auditing: Compare functional requirements with technical designs to ensure every user story has a corresponding technical implementation and database support.
  • Entitlement Verification: Compare monetization boundaries with functional flows to ensure free/paid tier limits are explicitly handled in requirements and UI screens.
  • Interface & Schema Validation: Compare OpenAPI contracts with database schemas to ensure data types, field names, and relational structures are perfectly synchronized.
  • Gherkin BDD Compliance: Compare modular Gherkin scenarios with PRD state transitions to ensure all edge cases and error paths are covered.

3. Consistency Audit Checklist

When performing an audit, you must systematically verify the following dimensions:

  1. PM vs. Architect (需求与架构对齐):
    • Are all features and state transitions in product/requirements.md and product/user-journey.md supported by architecture/architecture.md?
    • Are there any database tables or API endpoints defined in architecture/database-schema.sql or architecture/openapi.yaml that do not map to any product requirement?
  2. GTM vs. PM/UI-UX (商业与功能/交互对齐):
    • Are the free/paid tier boundaries and entitlement limits defined in product/monetization-plan.md explicitly reflected in product/requirements.md (e.g., limit checks, upgrade triggers)?
    • Does interface/ui-mockups.md include visual cues, paywalls, or upgrade buttons for paid-tier features?
  3. UI-UX vs. Architect (交互与技术对齐):
    • Are the interactive elements, forms, and data inputs shown in interface/ui-mockups.md fully supported by the API fields in architecture/openapi.yaml and columns in architecture/database-schema.sql?
    • Are the responsive breakpoints or multi-platform requirements supported by the architectural deployment model?
  4. Architect Schema vs. API (接口与数据库对齐):
    • Do the field names, data types, and primary/foreign key relationships in architecture/database-schema.sql match the request/response schemas in architecture/openapi.yaml?
  5. BDD vs. PRD:
    • Do the Gherkin BDD scenarios in modules/ cover all the Happy paths, edge cases, and error-handling flows defined in product/requirements.md?
  6. BR-DoD & Directory Structure Audit:
    • Audit whether all generated files are strictly stored under the five clear subdirectories (product/, interface/, architecture/, modules/, bridge/), avoiding root-level flat files, and ensuring no specs/ directory is generated.
    • Audit whether bridge/human-preparation.md is perfectly aligned with the external dependencies defined in product/monetization-plan.md and architecture/architecture.md. Are there any missing or redundant preparation items?
    • Audit whether the development sequence in bridge/ai-co-dev-guide.md strictly corresponds to the specifications in bridge/scaffolding-blueprint.md, architecture/, and modules/.
    • Audit whether the scaffolding directories and core code specifications in bridge/scaffolding-blueprint.md perfectly align with the technical choices and layered structure in architecture/architecture.md.
    • Ensure all external integrations support a complete sandbox/mock dual-track mode (if external services are designed).
    • Ensure standard user registration, login, and session recovery flows are fully designed (if authentication is designed).
    • Ensure system exceptions, offline states, or quota limits have corresponding UI-level feedback (Toasts/Modals) (Always Applicable).
    • Ensure a user-accessible global Settings Panel is designed and mounted (if configurable settings are designed).
    • Ensure the architecture includes unsigned/development build packaging and single-machine deployment configurations (where applicable based on the platform).

Read the full file on GitHub · 65 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. yesterday First seen · 65 lines · 61 tokens per session scan A 206765e6aeaf

Subscribe to this mod's changes

alignment-agent is an agent published in the GitHub repository Codoop/codoop-flow (5 stars, last pushed 8d ago), licensed MIT. It adds 61 tokens to every session and 1,515 once invoked, about $0.0003 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 2026-08-31.