Skip to content

AXIMO.md

Latest commit

 

History

History
58 lines (40 loc) · 2.92 KB

AXIMO.md

File metadata and controls

58 lines (40 loc) · 2.92 KB
Raw

Aximo Operating Instructions

Purpose

Aximo Control is the local operating control plane for coordinating projects, machines, agents, tasks, validations, and reports. This repository should remain Markdown/YAML/Python-first, private by default, and report-driven.

This file is the repo-level operating instruction file for any compatible assistant or coding tool. It is intentionally tool-agnostic: Claude Code, Codex, ChatGPT, Gemma Coder, OpenClaw, and future tools should operate through replaceable adapters rather than hard-coded assumptions.

Operating Model

  • Use Markdown for human-readable operating documents.
  • Use YAML for structured local state.
  • Use Python standard library scripts for local automation when needed.
  • Keep workflows file-based until there is a clear reason to add more infrastructure.
  • Prefer small, reviewable changes with explicit reports.
  • Keep commands, agents, skills, adapters, and hooks as documentation unless explicitly approved for implementation.

Safety Boundaries

  • Do not add secrets, tokens, credentials, private keys, customer data, production credentials, or sensitive infrastructure details.
  • Do not connect external services unless explicitly requested and approved.
  • Do not commit, push, create PRs, merge, deploy, publish, or change repository settings unless explicitly instructed.
  • Do not modify project repositories discovered by Aximo Control unless the current task explicitly targets that repository.
  • Public-facing, customer-facing, investor-facing, or partner-facing outputs require human approval.
  • Non-development department outputs are draft-only unless explicitly approved for external use by a human. This includes marketing, sales, customer support, docs/proposal, finance/admin, and other non-development operating outputs.

Workflow Entry Points

  • Start work with commands/start_task.md.
  • Close daily work with commands/eod.md.
  • Transfer context with commands/handoff.md.
  • Review operating health with commands/weekly_review.md.

Role Boundaries

  • Development implementation work should follow agents/dev_implementation_agent.md.
  • Validation work should follow agents/validation_agent.md.
  • Report creation should follow agents/report_agent.md.
  • Control-plane operations should follow agents/operations_agent.md.

Reusable Skills

  • Task packet creation: skills/task_packet_creation/SKILL.md
  • Validation reporting: skills/validation_report/SKILL.md
  • EOD reporting: skills/eod_report/SKILL.md
  • Git discovery: skills/git_discovery/SKILL.md

Tool Adapters

Adapters describe how tools should map into Aximo workflows without making Aximo depend on one tool:

  • ChatGPT: adapters/chatgpt.md
  • Codex: adapters/codex.md
  • Claude Code: adapters/claude_code.md

Planned Events

Future event names are documented in hooks/planned_events.md. They are not executable hooks and must not be treated as automation until explicitly implemented and approved.