Protocol Map: Find the Right Contract by Failure Mode
Choose the right AI Protocol Kit contract by failure mode. Six lanes cover ambiguity, wrong output, evidence loss, publication errors, and session drift.
Use this file to discover all available pages before exploring further.
This map is organized by failure mode, not by protocol name. Each lane describes a class of situation — what tends to go wrong, and why — and lists the protocols that address it. If you already know the failure mode you are trying to prevent, find that lane and use the protocol it points to. If you are not sure, scan the lane descriptions and pick the one that most closely matches what you are actually worried about in this specific task.
Shape the Work
Brief the Artifact
Review Findings & Evidence
Reasoning, Constraints & System Reading
GitHub / Public Repo & Page Publishing
Govern Agentic Sessions
Use these protocols when the task is still ambiguous, premature, overloaded, or underspecified. The common failure here is that AI collapses a fuzzy idea into a confident output too quickly — before the actual scope, constraints, or intent have been established. These protocols interrupt that collapse and force the shape of the work to be explicit before anything gets built or written.
Idea Shaping Protocol
Turn a rough idea into a clear structure before asking AI to write, plan, design, or build anything.
Pre-Task Expansion Protocol
Stop AI from collapsing too quickly into the obvious answer by exposing alternative readings, tensions, and surrounding shape first.
Structural Shaping Protocol
Shape ambiguous input into operational form before synthesis, with gates, evidence level, source boundaries, and artifact direction explicit.
System Reading Protocol
Read the gap between declared intent and observable behavior to extract the operative principle of a system without premature solutions.
Use these protocols when the AI needs a clear, complete brief before it starts writing, building, or structuring output. The failure mode here is output that is technically competent but wrong for the context — built for a generic reader, missing key constraints, or structured around assumptions that were never stated. These protocols close those gaps before a single line is drafted.
HTML Page & Tool Briefing Protocol
Guide AI from raw intake to a complete implementation brief for a static HTML page, web surface, or local tool.
Output for Real Readers Protocol
Write guides, README text, emails, pages, forms, or instructions with reader identity, purpose, tone, and language constraints closed before drafting.
Use these protocols when the job is inspection, bug capture, or evidence preservation — situations where the difference between what was observed and what was inferred matters. The failure mode is evidence drift: AI summarizes findings, the summaries become claims, and the claims are later treated as verified facts. These protocols enforce hard separation between observation, inference, uncertainty, and promotion.
Field Findings & Bugs Protocol
Capture findings and bugs as structured artifacts while keeping evidence, inference, uncertainty, relations, and later promotion separate.
System Reading Protocol
Read the gap between declared intent and observable behavior to extract the operative principle of a system without premature solutions.
Use these protocols when the problem is not just unclear but structurally prone to collapse — where AI might resolve a real tension with a false compromise, oversimplify a constraint conflict, or default to the most plausible-sounding answer rather than the most defensible one. These protocols require the AI to expand the problem space, surface competing interpretations, and hold conflicting constraints in view before producing output.
Pre-Task Expansion Protocol
Stop AI from collapsing too quickly into the obvious answer by exposing alternative readings, tensions, and surrounding shape first.
System Reading Protocol
Read the gap between declared intent and observable behavior to extract the operative principle of a system without premature solutions.
PHI-Lens Protocol
Handle non-trivial tasks where constraints interact and a flat compromise would hide the dominant force, counterforce, or asymmetry.
Use these protocols when the work touches GitHub repositories, README files, GitHub Pages, public project pages, discovery files, badges, or any publication step that is difficult to reverse. The failure modes here are publication without review, incorrect metadata, broken discovery files, and badges or counters added without a real purpose. These protocols add explicit confirmation gates before any commit, push, deploy, or public-facing change.
GitHub Repository Publication Protocol
Prepare a repository for clean GitHub publication with project classification, workspace audit, and confirmation before commits or push.
GitHub README Framing Protocol
Write or restructure a GitHub README from repository evidence, reader fit, and the real job the README must perform.
GitHub Pages Discovery Protocol
Prepare GitHub Pages sites with correct publication roots, discovery files, canonical URLs, sitemap, and robots handling.
GitHub Badge & Telemetry Protocol
Decide which badges, counters, analytics, or validation checks make sense for a repo without adding vanity widgets.
Public Page Publication Protocol
Prepare or publish a public page with page role, audience, metadata, accessibility, and publication risks closed before deploy.
HTML & Website Discovery Protocol
Prepare static HTML pages with machine-readable discovery files, head metadata, canonical URLs, sitemap, and robots rules.
Use these protocols when the AI session itself needs stronger behavioral rails — when the AI is operating across multiple files, tools, or repo states, or when a task requires multiple passes with review between them. The failure mode here is an AI that acts without grounding: writing files it has not read, confirming steps it has not verified, or continuing when the task is not actually ready for the next step. These protocols impose posture, ledger discipline, and executor/reviewer separation.
GPT Agentic Posture Contract
Make ChatGPT work more like Codex: tool-aware, grounded in real files, ledger-based, verification-oriented, and willing to stop when not ready.
Triad AI Orchestration Protocol
Run a review loop between an executing AI and a reviewing AI over snapshots, repo state, or diffs while preserving target authority.
Some protocols appear in more than one lane. System Reading Protocol appears under both “Shape the Work” and “Review Findings & Evidence.” Pre-Task Expansion appears under both “Shape the Work” and “Reasoning, Constraints & System Reading.” PHI-Lens appears under both “Reasoning, Constraints & System Reading” and “Shape the Work” contexts when constraint interaction is the dominant risk.This is intentional, not a mistake. These protocols are genuinely useful in more than one failure-mode context — they can play different roles depending on what the task needs. Seeing a protocol in two lanes does not mean you should load both lanes at once. It means the protocol is versatile, and your situation determines which lane’s framing is most relevant. Pick the lane that describes your actual failure risk, use the protocol it recommends, and do not stack unless you have a specific reason documented in the How to Use guide.
Canvas compatibility noteThe GPT Agentic Posture Contract depends on Canvas as a separate, persistent execution ledger. Since OpenAI’s May 28, 2026 GPT-5.5 update, Canvas is no longer available in GPT-5.5 Instant or GPT-5.5 Thinking. Without Canvas, the contract remains useful as posture guidance, but it should not be treated as a full agentic execution protocol.