Skip to content

Latest commit

 

History

History

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 

README.md

ai-instruction/ — briefing templates for large language models

Scope

This directory contains prompt templates and delegation guidance used when a human (or orchestrating model) gives work to a specific LLM model tier. It is advice to the prompter, not configuration consumed by a bot.

Three model tiers are covered, each in its own file:

  • haiku.md — mechanical sweeps, enumerable audits, formatting passes
  • sonnet.md — mid-tier implementation, non-trivial refactors, test authoring, local reasoning
  • opus.md — proof work, language/compiler design, novel architecture, cross-repo synthesis, supervision of other models

What this directory is NOT

This directory must not be confused with the estate's bot directive channel.

Concern Location Audience Format
Briefing templates for prompting LLMs (this dir) standards/ai-instruction/ Humans + orchestrating LLMs writing delegation prompts Markdown prose
Repo-local machine-readable directives <repo>/.machine_readable/ (6a2/, contractiles/, anchors/, etc.) gitbot-fleet, hypatia, coordination.k9, MCP guardian A2ML

The .machine_readable/ channel tells the gitbot-fleet how this particular repo behaves — its invariants, contractiles, neurosymbolic rules, canonical file locations. It is consumed mechanically by bots at CI time and by the MCP guardian at agent session start.

This directory (ai-instruction/) is a different channel entirely: it tells a prompter how to choose and structure a request to a given model tier so the output is useful. The bot fleet never reads these files; they are editorial guidance that lives alongside the other human-readable standards in this repo.

If you find yourself mixing the two — e.g. putting "tell Haiku to audit this repo" advice into a repo's .machine_readable/AGENTIC.a2ml — stop; that advice belongs here.

How to use

Before delegating a task to a model:

  1. Decide the tier (see the per-model files for task-fitness tables).
  2. Copy the relevant file's prompt scaffold and fill in the task-specific blocks.
  3. Include the model's hard rules section verbatim (the model will not follow rules it cannot see — memory and global CLAUDE.md do not transfer to delegated subagents).
  4. Decide your trust level ahead of time (per that model's trust guidance) and spot-check before acting.

Cross-references

License

PMPL-1.0-or-later (MPL-2.0 automatic legal fallback).