Appendix O
Book Index
A–Z topic index, every concept, tool, and named section across the manual with a direct link to the page that covers it.
How this is built. This page is auto-generated by
node docs/manual/maintain.mjs from the chapter list and curated section index in assets/manual.js. To add a new entry, add it to the relevant page and re-run the script. See also the Glossary for definitions of core terms.
A
A Day in the Forge — Three Vignettes
A Day in the Life of a Slice
About the Author
Actions Tab
Adoption Path - Two Routes (Stakeholder Briefing)
Advanced Execution
Agent Factory — The Recipe in One Page
Agent Factory Recipe
Agent Not Following Guardrails
Agents Don't Drift Maliciously
Air-Gapped Deployment
analyze vs diagnose
Anvil (L3 boundary, DLQ, capability handshake)
Anvil & Lattice Dashboard Tab
API Key Configuration
applyTo Pattern Reference
Audit Loop (Deep Dive)
Audit-Loop Activation
Auto-Loading Beats Manual
Azure Government
B
Bug Registry MCP Tools
C
Capacity Planning (Per-Team Sizing)
Check Prerequisites
Choosing Your Preset
CI Integration GitHub Actions
Claude Code Setup
CLI Reference
Clone and Run Setup
Cloud Agent
Codex Setup
Common Error Messages
Common Mistakes
Compliance — Audit Logging
Compliance — Data Flow
Compliance — Identity and Authentication
Compliance & Data Residency
Compliance Posture (SOC2 / HIPAA / PCI / FedRAMP / GDPR)
Config Tab
Configuration Hierarchy
Context Files per Slice
Conventions Used in This Manual
copilot-instructions.md
Core MCP Tools
Cost — Anti-lock-in posture (BYOK, no proxy, no telemetry, open pricing)
Cost — Cost drivers (model tier, tokens, quorum, cache, reasoning, retries)
Cost — Cost-effective workflows (slice sizing, routing, gates, cache, quorum)
Cost — Estimate vs actuals (forge_estimate_quorum vs forge_cost_report)
Cost — Forecasting at scale (groupBy model / role / scope)
Cost — Orientation (BYOK, no markup, per-slice attribution)
Cost — Per-quorum-mode economics (auto / power / speed / disabled)
Cost — Three sources of truth (pricing table, estimators, actuals)
Cost — Worked example (slice B5 ship REST API reference)
Cost & Economics
Cost Discipline
Cost Optimization
Cost Tab
Cost Tracking
Costs Are Too High
Creating Extensions
Cross-Stack Agents
Crucible (Idea Smelting)
Crucible MCP Tools
Cursor Setup
Custom Instruction Files
Customization
D
Dashboard — Forge-Master
Dashboard — LiveGuard
Dashboard — Settings
Dashboard Won't Load
Day 1 — Pilot Installation
Diagnostic Tools
Discovery Harness Implementation
Domain Instruction Files
E
Easy Button (one-prompt install)
End-to-End Workflow: WhatsApp to Shipped PR
Enterprise Architect Ladder (Reader Paths)
Enterprise Reference Architecture
Env Vars — Azure OpenAI Alternative Routing
Env Vars — CLI Internal (set transiently by pforge)
Env Vars — Feature Toggles
Env Vars — Host Detection (read-only)
Env Vars — Orchestrator Timing (gate, worker timeouts)
Env Vars — Project and Runtime
Env Vars — Provider API Keys (XAI, OpenAI, Anthropic)
Env Vars — Resolution Precedence
Env Vars — Server Ports and Network
Env Vars — Telemetry (OpenTelemetry)
Env Vars — Worked Example (PowerShell profile)
Env Vars Reference — Orientation
Environment Variables Reference
Errors & Exit Codes
Errors & Exit Codes — CI / scripting recipes
Errors & Exit Codes — Error events on the hub
Errors & Exit Codes — MCP tool errors (forge_* envelope)
Errors & Exit Codes — Named error catalog (A-Z)
Errors & Exit Codes — Orchestrator exit codes & statusReason
Errors & Exit Codes — Orientation (4 layers)
Errors & Exit Codes — OS subprocess exits (Ctrl+C, SIGKILL, SIGTERM)
Errors & Exit Codes — pforge CLI exit codes (0/1/2)
Errors & Exit Codes — REST error shape (HTTP 400/404/409/429/500)
Escalation Chains
Estimating Quorum Cost forge_estimate_quorum
Event Catalog
Event Catalog — Bridge (approval-*, bridge-notification-*)
Event Catalog — Client→server (set-label)
Event Catalog — Common Envelope (version, type, source, security_risk)
Event Catalog — Consuming the Stream (WebSocket subscription)
Event Catalog — Crucible (crucible-smelt-*)
Event Catalog — Escalation & CI (slice-escalated, ci-triggered)
Event Catalog — Lifecycle (run-started, slice-*, run-completed)
Event Catalog — LiveGuard (drift, incident, secret-scan, watch-*)
Event Catalog — Orientation
Event Catalog — Retention (hub ring, run journal, LiveGuard cache, OpenClaw)
Event Catalog — security_risk enum
Event Catalog — Skills (skill-started, skill-step-*)
Event Catalog — source enum
Event Catalog — Tempering (bug-validated-fixed)
Evidence A/B Test Results
Execute the Plan (Quickstart)
Executive Summary (Stakeholder Briefing)
Extension Author Ladder (Reader Paths)
Extension Catalog
Extensions
F
.forge.json — agents (claude, cursor, codex)
.forge.json — brain.federation (cross-project memory)
.forge.json — Execution Limits (parallelism, retries)
.forge.json — extensions
.forge.json — forgeMaster reasoning loop
.forge.json — Full Annotated Example
.forge.json — hooks.postSlice (drift thresholds)
.forge.json — hooks.preAgentHandoff
.forge.json — hooks.preDeploy (LiveGuard)
.forge.json — meta.selfRepairRepo
.forge.json — modelRouting (default, execute, review)
.forge.json — openclaw analytics bridge
.forge.json — Project Identity (projectName, preset)
.forge.json — quorum (multi-model consensus)
.forge.json — runtime.gateSynthesis (Phase-25 L6)
.forge.json — runtime.reviewer (Phase-25 L4)
.forge.json — testbed.path
.forge.json — updateSource (auto / github-tags)
.forge.json Config
.forge.json Reference
.forge.json Reference — Orientation
Failure Mode FM1 — Token limit hit
Failure Mode FM10 — Worker spawn failure
Failure Mode FM11 — Git stash conflict on rollback
Failure Mode FM12 — Snapshot apply failure
Failure Mode FM13 — Plan parse error
Failure Mode FM14 — Provider rate limit (HTTP 429)
Failure Mode FM15 — Provider 5xx / outage
Failure Mode FM16 — Auth expired
Failure Mode FM17 — L2 jsonl corruption
Failure Mode FM18 — L3 endpoint unreachable
Failure Mode FM19 — Hook false positive
Failure Mode FM2 — Model timeout
Failure Mode FM20 — Hook script error
Failure Mode FM21 — Quorum panel disagrees below threshold
Failure Mode FM22 — Quorum panelist timeout
Failure Mode FM23 — Port already in use
Failure Mode FM24 — Disk full
Failure Mode FM25 — File locked (Windows)
Failure Mode FM3 — Malformed tool call
Failure Mode FM4 — Edit blocked by scope / forbidden actions
Failure Mode FM5 — Worker loop detected
Failure Mode FM6 — Gate test failure (legitimate)
Failure Mode FM7 — Gate timeout
Failure Mode FM8 — Non-portable gate command
Failure Mode FM9 — Documentation validator drift
Failure Modes — General recovery techniques
Failure Modes — Index (25 failure modes across 8 layers)
Failure-Mode Catalog
Feature Parity Matrix
Fleet KPIs
Fleet Operator Playbook
Focused Instructions Beat Generic Ones
Foreword — From Impossible to Seven Minutes
forge_abort Stop Execution
forge_analyze Consistency Scoring
forge_capabilities Discovery
forge_diagnose Bug Investigation
forge_estimate_quorum Cost Preview
forge_generate_image
forge_plan_status Execution Status
forge_run_plan Execute Plan
forge_smith Environment Check
forge_sync_memories (Copilot Memory soft-sync)
Forge-Master
Forge-Master MCP Tool
Forge-Master Studio Tab
G
Gemini Setup
Generic Enterprise Reference Architecture
GitHub Stack Alignment
Glossary
Grok Image Generation
Grok Image Warnings
H
Hallmark (provenance, hallmark/v1)
Harden the Plan (Quickstart)
Health DNA
Host-Aware Routing
How Do I — Brief Stakeholders and Onboard Readers
How Do I — Customize Plan Forge for My Project
How Do I — Debug and Troubleshoot
How Do I — Execute a Plan
How Do I — Extend and Integrate
How Do I — Install and Set Up
How Do I — Operate at Scale (Teams and Fleets)
How Do I — Plan a Feature
How Do I — Review and Ship
How Do I — The Nine Intent Groups
How Do I…? — Task Index
How Guardrails Auto-Load (applyTo)
How It Works
How Plan Forge Composes with GitHub
How the New Memory Pieces Fit the Old Tiers
How the Shop Remembers
How To Read This Book (Foreword)
I
Independent Review (Quickstart)
Independent Review Catches What Builds Miss
Install
Installation
Installing Extensions
Instruction Files & Agents
Instructions & Agents — Reference
Integrating from Outside
L
Lattice (code-graph, chunker, callers, blast)
Lessons Learned
Lifecycle Hooks
Lifecycle Hooks — Copilot session (SessionStart, PreToolUse, PostToolUse, Stop)
Lifecycle Hooks — LiveGuard (PreDeploy, PostSlice, PreAgentHandoff)
Lifecycle Hooks — Plan-execution guard (PreCommit)
Lifecycle Hooks — Resolution order
Lifecycle Hooks — Writing a custom hook
Lifecycle Hooks Reference — all eight hooks
List of Figures
LiveGuard Alert Runbooks
LiveGuard Env Tab
LiveGuard Health Tab
LiveGuard Incidents Tab
LiveGuard MCP Tools
LiveGuard Security Tab
LiveGuard Tools Reference
LiveGuard Triage Tab
M
Make This Yours - Tailoring Flow (Stakeholder Briefing)
MCP Server — Quick Start
MCP Server — Reference
MCP Server & Tools
MCP Server Architecture
MCP Server Chapter Overview
MCP Server Selection (Plan Forge / GitHub / Foundry Toolbox)
MCP Tools 69 Categories
Memory Architecture
Microsoft Foundry Composition Variant
Model Routing
Multi-Agent Quorum Turns PFORGE_QUORUM_TURN
Multi-Agent Setup
Multi-Agent Setup
Multi-Team Operations (Federated vs Centralized)
N
Nested Subagents
Network and Isolation Patterns (Cloud / Hybrid / Air-Gapped)
O
Observability Export (OTel)
One-Click Install
OpenBrain Memory
OpenBrain: The Connective Tissue
OTLP Telemetry Traces
P
Parallel Execution [P] tag
Parallel Execution DAG
pforge analyze
pforge check
pforge diagnose
pforge diff
pforge ext
pforge init
pforge run-plan
pforge smith
pforge smith Verification
pforge status
pforge sweep
pforge update
Pick Your Preset
Pipeline Agents
Pipeline Agents Click-Through
Plan Execution Fails
Plan Forge for Enterprise
Plan Forge on the GitHub Stack
Plan Pattern Library
Plan Pattern P1 — Add an Entity (DB → service → API → UI)
Plan Pattern P10 — Performance Fix (benchmark-driven)
Plan Pattern P11 — Security Patch (CVE / vulnerability)
Plan Pattern P12 — Documentation Phase (one slice per document)
Plan Pattern P13 — CI/CD Workflow Change (no-op + promote)
Plan Pattern P14 — Spike-Then-Build (time-boxed exploration)
Plan Pattern P2 — Add an Endpoint (new route on existing entity)
Plan Pattern P3 — Add an External Integration (third-party API)
Plan Pattern P4 — Refactor a Subsystem (multi-consumer migration)
Plan Pattern P5 — Fix a Regression (strict red-green-refactor)
Plan Pattern P6 — Hotfix (minimal-surface emergency change)
Plan Pattern P7 — Feature Flag Rollout (ship dark, toggle later)
Plan Pattern P8 — Data Migration (additive + backfill + verify)
Plan Pattern P9 — Dependency Upgrade (per-module fix slices)
Plan Patterns — Anti-patterns (mega-slice, test-after, etc.)
Plan Patterns — Composing patterns across phases
Plan Patterns — Index of 14 patterns (when, slice count)
Plan Structure
Plans Are Markdown
Pre-flight Check (Quickstart)
Prerequisites
Progress Tab
Project History
Project Principles
Project Profile
Publishing Extensions
Q
Quick Reference Card
Quick Start for Evaluators
Quorum Complexity Scoring Rubric
Quorum Mode
Quorum Mode in Practice (Day in the Forge)
Quorum Quality Examples - 3 Models vs 1
Quorum vs Quorum Advisory
R
Reader-Journey Ladders — Pick Your Path
Reading the Hardened Plan
Replay Tab
REST API — Authentication, binding, and CORS
REST API — Bridge and approvals
REST API — Copilot integration
REST API — Cost
REST API — Crucible (idea smelting)
REST API — Discovery (well-known, capabilities, version, status)
REST API — Error response shape
REST API — Forge-Master (conversational entrypoint)
REST API — Generic MCP dispatcher (POST /api/tool/:name)
REST API — GitHub and team coordination
REST API — Image generation
REST API — Inner loop (reviewer calibration, gate suggestions)
REST API — LiveGuard (drift, incidents, deploys, secret scan)
REST API — Memory (L1/L2/L3)
REST API — Notifications, audit, dashboard, settings
REST API — Orientation (16 subsystems, 113 endpoints)
REST API — Plan execution and runs
REST API — Quorum and fix proposals
REST API — Search, timeline, hub
REST API — Skills (decision tray)
REST API — Tempering and bugs
REST API — Worked Examples (curl, wscat, SDK)
REST API Endpoints
REST API Reference
Resume and Retry
Review & Ship
Reviewer or Architect Ladder (Reader Paths)
Runs Tab
S
Sample Project
Scaling the Factory Across Teams
SDK for Integrators
Security — AI-specific threats (prompt injection, untrusted tool output, scope escape)
Security — Attack surface enumeration
Security — Hardening checklist (12 controls)
Security — Incident response (LiveGuard front door)
Security — Orientation (developer-machine-first posture)
Security — Prompt injection defenses
Security — Sandboxing & gate execution (TCB boundary)
Security — Scope escape (drift detection, Review Gate)
Security — Secret management (env, .forge/secrets.json, gh auth)
Security — STRIDE per subsystem
Security — Supply chain (Plan Forge itself, extensions, providers)
Security — Trust boundaries (6 boundaries)
Security — Untrusted tool output defenses
Security & Threat Model
Self-Deterministic Loop (Deep Dive)
Sessions and Why They Matter
Settings API Keys Tab
Settings Brain Tab
Settings Bridge Tab
Settings Crucible Tab
Settings Execution Tab
Settings General Tab
Settings Memory Tab
Settings Models Tab
Settings Updates Tab
Setup Failed
Setup Wizard
Ship (Quickstart)
Skills — Authoring a New Skill
Skills — Events Emitted by the Runner
Skills — Shared Skills (every preset)
Skills — SKILL.md Runtime Contract
Skills — Stack-Specific Skills (per language)
Skills — Three Ways to Invoke
Skills Slash Commands
Skills Tab
Slice Boundaries Matter More Than You Think
Slices Gates and Scope
Slicing Strategy
Solo Developer Ladder (Reader Paths)
Spec Kit Ecosystem Extensions
Spec Kit Import Flow
Spec Kit Import Procedure
Spec Kit Interop
Spec Kit Interop
Specify the Feature (Quickstart)
Stack-Specific Agents
Stack-Specific Notes
Stakeholder Briefing — the 10-minute white paper
Starting the Dashboard
Starting the MCP Server
Step 0 Specify the Feature
Step 2 Harden the Plan
Step 3 — Route Agents to Lanes
Step 3 Execute
Step 5 Review
Stop Conditions
Studio Classification Badge
Studio Quorum Advisory
Studio Session Persistence
Sweep for Deferred Work (Quickstart)
T
Team Coordination
Team Lead Ladder (Reader Paths)
Tempering MCP Tools
Testbed MCP Tools
The .NET A/B Test — 99 vs 44 (Day in the Forge)
The 7-Step Pipeline
The 80/20 Wall
The Blacksmith Analogy
The Bug Registry
The Competitive Loop (Deep Dive)
The Compounding Flywheel (Stakeholder Briefing)
The Copilot Integration Trilogy
The Dashboard
The File System
The Five Ladders at a Glance (Reader Paths)
The Four Cost Levers (Stakeholder Briefing)
The Four New Pieces (Hallmark, Anvil, Lattice, sync_memories)
The Four-Station Shop (Foreword)
The Inner Loop (Deep Dive)
The Knowledge Graph
The LiveGuard Dashboard
The Loop That Never Ends (Day in the Forge)
The One-Paragraph Version (Foreword)
The Problem in One Sentence
The Remote Bridge
The Testbed
The Watcher
Three Memory Commands You Can Run Today
Three Vignettes at a Glance (Day in the Forge)
Three-Lane Triage Funnel
Timeline Tab
Traces Tab OTLP
Troubleshooting
Troubleshooting — Errors & Exit Codes quick reference
Two-Layer Guardrail Model
Typical MCP Workflow
U
Unified API Surface Index
Unified Memory Across Agents
Universal Instruction Files
Update Source Modes
Updating Plan Forge
V
v1.0 Foundation
v2.0 Autonomous
v2.10 OpenClaw
v2.14 GitHub Copilot Integration
v2.18 Temper Guards
v2.5 Quorum Mode
v2.83 Host-Aware Routing
v2.95 Lattice / Code-Graph
v3.0 Copilot Trilogy
v3.2–3.4 Team Mode
v3.6 OpenBrain L3 (current)
Validation Gates
Verify MCP Server Running
Verify with pforge smith
W
Watcher MCP Tools
Watcher Tab
WebSocket Hub Events
Week 12 — Full Fleet Quarterly Review
Week 4 — Pilot Graduation
What Changed (and What Did Not)
What GitHub Leaves to the Ecosystem
What GitHub Ships (the Substrate)
What Happens Without Guardrails
What Is LiveGuard?
What Is Plan Forge?
What Plan Forge Does
What Plan Forge Is and Is Not (Stakeholder Briefing)
What the Three Vignettes Share (Day in the Forge)
What This Book Is Not (Foreword)
What This Is Not
What We Add You Didn't Ask For (Stakeholder Briefing)
Whats Next After Quickstart
When Two Ladders Apply (Reader Paths)
Where to Find What You Need (Enterprise)
Who This Is For
Why Cheaper Models Punch Above Their Weight
Why Open Source Matters (Stakeholder Briefing)
Why Plan Forge for the Enterprise
Why Session Isolation Works
Windsurf Setup
Worked Example - Copilot CLI + Grok API
Writing a Good Scope Contract
Writing Plans That Work
Y
Your First Plan
Your First Plan