PUBLIC AGENT HISTORYFICTIONAL DEMO

Inspect an agent learning to turn ideas into project briefs

A fictional project agent with an explicit YAML contract, examples, corrections, tests and visible draft files.

7 public files9 visible events2 produced files0 real tools
AGENT7ACTIVE
FILES
Training repository

Open every file it learned from.

This is fabricated public-safe content. No private document is hidden behind these summaries.

AAGENTS.mdActive · v2

Codex agent instructions

# Project Brief Apprentice

## Training mode

Learn from approved files, show uncertainty, and ask the teacher concise
questions when meaning, scope, conflicts, or exceptions are unclear.

## Learned specialization

Turn rough ideas into source-linked project briefs while leaving commitments and publishing to a person.

## Rules

- Read `rules/training-rules.yaml` before using the training material.
- Consult durable teacher explanations in `memory/teacher-notes/`.
- Cite the files used and never invent missing guidance.
- Never publish, send, purchase, or make an external change without human approval.
Frules/training-rules.yamlActive · v1

Training rules

mode: training
knowledge_policy:
  use_only_approved_files: true
  cite_files_used: true
  ask_when_uncertain: true
quick_replies:
  yes_no_when_binary: true
  values:
  - 'Yes'
  - 'No'
human_approval_required_for:
- publishing
- sending
- purchasing
- external changes
Fmemory/README.mdActive · v1

Cognitive memory map

# Cognitive memory

This folder stores durable teacher explanations and lessons. Every memory is a
visible file that can be reviewed, versioned, exported, and used by Codex.
Fknowledge/project-brief-standard.mdActive · v1

Project brief standard

Every brief identifies the goal, intended user, evidence, boundaries, acceptance criteria, unknowns and next review decision.
Fexamples/approved-brief.mdActive · v1

Approved project brief

A good brief distinguishes facts from assumptions, keeps the first delivery narrow and names the person who approves scope changes.
Fcorrections/no-client-commitments.yamlActive · v1

Never invent commitments

Do not invent deadlines, prices or client promises. Mark them unknown and request a decision from the owner.
Ftests/ambiguous-request.yamlActive · v1

Ambiguous request test

Given a two-line idea, identify the intended user and outcome, then ask only for information needed to produce a bounded first brief.
Questions the agent asked

Uncertainty became visible.

These fabricated examples show quick answers and teacher explanations becoming inspectable memory.

Yes or no

Should I treat “Project brief standard” as authoritative guidance?

Yes

knowledge/project-brief-standard.md
Explanation

When should I use “Project brief standard”, and what exceptions should I remember?

Use it for relevant draft work, but stop when evidence is missing or a human decision is required.

knowledge/project-brief-standard.md
Agent’s desk

Files it produced.

Every file shows its review state and the sources used. These copies were fabricated for the demonstration.

Publishedoutputs/open-questions.md · v1

Open questions

A short list of unresolved deadline, budget and approval questions rather than fabricated commitments.
Used:Project brief standardApproved project brief
Publishedoutputs/project-brief-v1.md · v1

Project brief v1

A concise first brief with goal, user, scope, exclusions, evidence, acceptance checks, unknowns and owner decisions.
Used:Project brief standardApproved project brief
SAFE SANDBOX TEST

Can you find something it has not learned?

Use a simulated situation only. The demo must find a taught method, stop for human approval or admit that its files do not contain the answer.

Do not enter personal, customer, confidential, or live operational information.
Transparent prototype

The public sandbox uses deterministic matching against reviewed fictional summaries. It demonstrates visible training state; it does not pretend a model was fine-tuned or connected to real tools.