AGENT MANIFESTO

How To Use

The framework is delivered as a set of prompt files. Pick the one that matches your situation and attach it to your AI tool.

version: 2.10.0
project: agent-manifest
url: https://github.com/AlexeyPlatkovsky/agent-manifest/blob/main/README.md

AI Instruction Framework

The Agent Manifesto is a portable, tool-agnostic framework for organizing AI instruction systems. MANIFEST.md defines the framework's values and principles; IMPLEMENTATION.md defines the practices that apply them. Together they keep instruction systems minimal, explicit, and scalable across single-tool and multi-tool environments.

How To Use

The framework is delivered as a set of stages. A stage is any NN_name.md file. Each stage is a self-contained entry point with a declared required context list — attach or reference it in your AI tool and ask the tool to run it. The exact syntax depends on your tool (@file in Claude Code, Cursor, and most modern agents), but the idea is the same across all of them:

run @<stage-file>.md

Run 00_project_profile.md before any other stage. After that, pick the stage that matches your current situation.

Stage 00 — Profile The Project

When: before any other framework stage, or when the user's role, duties, tools, or project assumptions changed.

Run:

run @00_project_profile.md

What happens:

  • The AI captures the project purpose, user role, recurring duties, and AI tool surface.
  • It identifies authoritative local sources, domain vocabulary, and quality expectations.
  • It optionally researches current best practices when local context is insufficient and the user approves.

Outcome: .ai/docs/project_specification.md, the reusable profile required by every later stage.

Stage 01 — Compose The Initial System

When: starting from scratch, or refactoring an existing messy instruction system.

Run:

run @01_initial_composition.md

What happens:

  • The AI inventories your repository.
  • It reads .ai/docs/project_specification.md.
  • It asks only unresolved design questions needed for composition.
  • It derives required capabilities from protocol metadata.
  • It derives required agents from agent template metadata.
  • It makes tool-specific adapters explicit enough to enforce the canonical root contract.
  • It checks repeated software work such as feature implementation, code review, and code refactoring as pipeline candidates when their steps differ.
  • It preserves good existing capability names where they already satisfy the framework.
  • It asks before any risky change (splits, moves, merges, deletions, contract choices).

Outcome: the smallest coherent instruction system that fully aligns with MANIFEST.md and IMPLEMENTATION.md.

Stage 02 — Review For Compliance

When: after significant instruction changes, or when you want a compliance check on an existing system.

Run:

run @02_review.md

What happens:

  • Validates the correct root-contract model.
  • Checks routing gates, duplication, and responsibility boundaries.
  • Verifies protocol coverage from structured metadata.
  • Verifies required agent coverage from structured metadata.
  • Produces a minimal fix plan before any implementation.

Outcome: a compliance verdict and minimum fix plan; fixes are implemented only when explicitly requested.

Stage 03 — Expand Capabilities

When: a valid baseline already exists and the team has real recurring workflows to encode.

Run:

run @03_capability_expansion.md

What happens:

  • Learns recurring work directly from you.
  • Proposes new skills, pipelines, agents, conventions, and docs grounded in actual usage.
  • Verifies present mandatory protocol triggers and mandatory agent-template triggers before optional additions.

Outcome: an instruction system that reflects how your team actually works, without speculative abstractions.

Stage 04 — Adopt External Tools

When: adopting a specific external tool, library, or framework into an existing instruction system after a valid baseline exists.

Run Stage 03 first if the tool introduces new capability triggers.

Run:

run @04_tool_adoption.md

What happens:

  • Inventories the tool's runtime surface, demos, and foreign instruction artifacts.
  • Reconciles foreign capabilities into standalone project artifacts, wrapped libraries, references, or discards.
  • Enforces cleanup of demo content and broken imports before completion.

Outcome: the external tool is cleanly integrated, with no leftover demo noise or conflicting instructions.

What This Repository Contains

  • MANIFEST.md — framework values and principles
  • IMPLEMENTATION.md — framework mechanics and operational rules
  • conventions/*.md — shared framework standards used by multiple framework artifacts
  • protocols/_README.md — protocol index
  • protocols/*.md — canonical protocol definitions used by stages
  • agents/*.md — canonical agent templates copied into generated landscapes when their metadata applies (agents/_README.md is the index)
  • 00_project_profile.md — creates or updates the reusable project specification
  • 01_initial_composition.md — builds or adjusts a baseline instruction system
  • 02_review.md — audits an instruction system against the framework
  • 03_capability_expansion.md — expands a correct baseline with new capabilities
  • 04_tool_adoption.md — adopts an external tool or framework into an existing instruction system

License

© Alexey Platkovsky. Licensed under CC BY 4.0.