An open wooden task-board mounted on the stone wall of the Plan Forge shop, a pegboard with dozens of small parchment task cards hung in tidy numbered rows and columns, glowing amber priority tabs attached to some cards, a smith's calloused hand reaching to pull one card down for the day's work

Appendix S

How Do I…? — Task Index

A task-first index over the rest of the manual. Find the verb that matches what you are trying to do; follow the link to the chapter that owns the answer. This appendix adds no new prose, it is pure navigation, sorted by intent rather than by where the chapters happen to live in the book.

Audience: Anyone who knows what they need to do but is not sure which chapter to open. Especially useful when returning to the manual mid-task.

How to use: Pick the intent group closest to your situation, scan the questions, click the answer link. If a task spans multiple chapters, the index lists each cross-ref, read them in order. If you cannot find what you need here, the Book Index covers concepts and the search box in the sidebar covers everything else.

The nine intent groups

The index is organized by what you are doing, not by what part of Plan Forge you are touching. Most tasks pull in two or three chapters across different Parts:

Intent group	When to use
1. Install & set up	You are putting Plan Forge on a fresh machine, or onto a new repository.
2. Plan a feature	You have a feature in mind and need to turn it into a hardened plan the Forge can execute.
3. Execute a plan	The plan exists; you are about to (or are mid-way through) running it.
4. Review & ship	The slices have run; you are deciding whether to merge and what to do post-merge.
5. Customize Plan Forge for my project	You want the agent to follow your team's specific patterns, not just the defaults.
6. Operate at scale (teams & fleets)	You are running Plan Forge across multiple repositories, multiple teams, or in an enterprise context.
7. Debug & troubleshoot	Something is broken, missing, or behaving unexpectedly.
8. Extend & integrate	You want to add new tools, glue Plan Forge to your existing systems, or build something on top.
9. Brief stakeholders & onboard readers	You need to walk a colleague, manager, or VP through what Plan Forge is and why it matters.

1. Install & set up

How do I install Plan Forge from scratch? Quickstart Step 1 — Install; then check prerequisites, clone & run setup, pick a preset, and verify with pforge smith.
How do I install Plan Forge with a single prompt? Quickstart — the Easy Button (one-prompt install).
How do I pick the right preset for my stack? Installation — Choosing Your Preset; then Customization — Project Profile if no preset fits exactly.
How do I verify Plan Forge is wired up correctly? Installation — Verify with pforge smith; expected output is mirrored in Troubleshooting — What a healthy pforge smith looks like.
How do I install for a different AI agent (Claude Code, Cursor, Codex, Gemini, Windsurf)? Chapter 13 — Multi-Agent Setup; per-agent sections: Claude Code, Cursor, Codex CLI, Gemini CLI, Windsurf, generic AI tool.
How do I update Plan Forge after install? Installation — Updating; and Appendix G — Update Source Modes for managed-vs-local choices.

2. Plan a feature

How do I take a fuzzy idea and turn it into a plan? Chapter 5 — Crucible (Idea Smelting); especially the Interview Loop and the three lanes (bug / Crucible / noise).
How do I write a plan that actually executes cleanly? Chapter 4 — Writing Plans That Work; start with the Scope Contract, then slicing strategy, validation gates, and stop conditions.
How do I avoid the most common plan-writing mistakes? Writing Plans — Common Mistakes.
How do I make sure my plan has the fields the executor needs? Writing Plans — CRITICAL_FIELDS Gate; the same gate fires in the Crucible.
How do I find a plan template for my situation? Writing Plans — Plan Templates.
How do I structure context files so the agent has what it needs? Writing Plans — Context Files.
How do I use Spec Kit specs alongside Plan Forge? Spec Kit Interop; coexistence rules are in Crucible — Spec Kit Coexistence.

3. Execute a plan

How do I run my first plan end-to-end? Quickstart Step 2 — Your First Plan for the 30-minute path; Chapter 6 — Your First Plan for the deeper walk-through.
How do I see what the agent is doing while it runs? Dashboard — Progress Tab; the Runs Tab for past runs and the Replay Tab for step-by-step replay.
How do I track how much a run is costing me? Dashboard — Cost Tab; for forward-looking estimates, see Advanced Execution — Estimating Quorum Cost.
How do I lower the cost without losing quality? Advanced Execution — Cost Optimization; pair with Model Routing and the Memory System discussion of cheaper models punching above their weight.
How do I get higher quality on a risky slice? Advanced Execution — Quorum Mode; for what changes in the output, see Quorum Quality Examples and Appendix R Vignette 3 — Quorum Mode in Practice.
How do I decide when quorum is worth the extra cost? Advanced Execution — Complexity Scoring Rubric (the seven signals that drive --quorum=auto).
How do I resume a plan after it failed mid-run? Advanced Execution — Resume and Retry.
How do I run multiple slices in parallel? Advanced Execution — Parallel Execution; plan-side prerequisites are in Writing Plans — Parallel Execution.
How do I run a plan as a cloud agent rather than locally? Advanced Execution — Cloud Agent Execution.

4. Review & ship

How do I review what the Forge produced before merging? Quickstart Step 3 — Review & Ship.
How do I detect deferred work, stubs, or TODOs? Quickstart — Sweep for Deferred Work.
How do I trust an independent reviewer with the result? Quickstart Step 3 covers the gate; the gate machinery is in Chapter 2 — How It Works.
How do I run LiveGuard before deploying so the deploy itself is safe? What Is LiveGuard? — When to Run LiveGuard Tools; tool surface in Chapter 17 — LiveGuard Tools Reference.
How do I follow up when LiveGuard finds something post-deploy? Appendix F — LiveGuard Alert Runbooks.

5. Customize Plan Forge for my project

How do I tell the agent what my project's rules are? Customization — Project Principles; for stack-specific rules, Project Profile.
How do I make the agent follow my coding standards? Customization — Writing Custom Instruction Files; how the agent picks them up is in Instruction Files & Agents — How Auto-Loading Works.
How do I see exactly which instruction files load for a given file? Instructions & Agents — A Concrete Scenario; reference layout in Instructions & Agents — Reference.
How do I add a custom reviewer or pipeline agent? Customization — Customizing Agents.
How do I add a custom slash-command skill? Customization — Customizing Skills.
How do I understand which file overrides which when settings conflict? Customization — Configuration Hierarchy.
How do I edit copilot-instructions.md safely? Customization — Editing copilot-instructions.md.
How do I configure .forge.json? Appendix T — .forge.json Reference; every key with type, default, example, and change impact.
How do I set provider API keys (XAI, OpenAI, Anthropic) safely? Appendix U — Provider API Keys; prefer .forge/secrets.json over committing keys to .forge.json.
How do I move Plan Forge off the default ports 3100 / 3101? Appendix U — Server Ports.

6. Operate at scale (teams & fleets)

How do I coordinate Plan Forge across multiple developers on one team? Chapter 27 — Team Coordination; especially the activity ledger and forge_team_dashboard.
How do I run Plan Forge across multiple teams or product groups? Appendix M — Fleet Operator Playbook; calendar-paced from Day 0 through Week 12.
How do I roll Plan Forge out to an enterprise? Appendix J — Plan Forge for Enterprise; start at Quick start for evaluators; reference deployments in Appendix K — Enterprise Reference Architecture.
How do I build a standardized agent fleet for my organization? Appendix L — Agent Factory Recipe; the seven steps from substrate check through iterate.
How do I attribute cost across teams? Fleet Operator Playbook — Cost attribution; per-run cost lives in Dashboard Cost Tab.
How do I keep multi-team operations from stepping on each other? Fleet Operator Playbook — Multi-team operations; pick federated or centralized.
How do I meet compliance and data-residency requirements? Appendix N — Compliance & Data Residency.

7. Debug & troubleshoot

How do I diagnose why a plan failed? Troubleshooting — Plan Execution Fails; start from the diagnostic tools overview.
How do I figure out why the agent is ignoring my guardrails? Troubleshooting — Agent Isn't Following Guardrails.
How do I fix a dashboard that won't load? Troubleshooting — Dashboard Won't Load; port reference is in Dashboard — Port Reference.
How do I read a cryptic error message? Troubleshooting — Common Error Messages.
How do I diagnose why setup itself failed? Troubleshooting — Setup Failed.
How do I diagnose why costs are higher than expected? Troubleshooting — Costs Are Too High; pair with Cost Optimization.
How do I diagnose why the Crucible can't finalize a smelt? Troubleshooting — Crucible Finalize Fails; Crucible-side detail in Crucible — Troubleshooting.
How do I diagnose why Forge-Master picks the wrong tool? Troubleshooting — Forge-Master Misroutes Intent.
How do I diagnose host-aware routing confusion? Troubleshooting — Host-Aware Routing Confusion; the design itself is in Advanced Execution — Host-Aware Routing.
How do I look at a known-good run and learn from it? Chapter 24 — The Testbed; the learn-by-doing reference testbed.

8. Extend & integrate

How do I browse what extensions are available? Chapter 12 — Browsing the Catalog; featured extensions for the curated set.
How do I install an extension? Extensions — Installing an Extension; later management in Managing Extensions.
How do I write my own extension? Extensions — Creating Your Own Extension; then publish it.
How do I drive Plan Forge from a script? Integrating from Outside — CLI; full command reference in Chapter 8 — CLI Reference.
How do I call Plan Forge from a service I am building? Integrating from Outside — REST API for request/response; WebSocket hub for live events; pforge-sdk for Node.js clients.
How do I pick the right integration surface? Integrating from Outside — Picking the right surface; auth in Auth and secrets.
How do I query the codebase as a graph? Knowledge Graph — forge_graph_query; pattern detectors in forge_patterns_list.
How do I sync instructions and memories with GitHub Copilot? Chapter 26 — Copilot Integration Trilogy; forge_sync_instructions for always-true rules; forge_sync_memories for learned lessons; when to run what.
How do I expose Plan Forge tools to my MCP-aware editor? Chapter 11 — MCP Server & Tools; MCP Server — Quick Start to wire it up.

9. Brief stakeholders & onboard readers

How do I brief a manager or VP on Plan Forge in one sitting? Stakeholder Briefing; the 10–15 minute white-paper version with all the canonical numbers.
How do I tailor a briefing to my own organization? Stakeholder Briefing — Make This Yours / Tailoring Flow; pick the template path, the slash-command-skill path, or the community path.
How do I onboard a new developer to my Plan-Forge-enabled project? route them to the Solo Developer Ladder if they are working alone, or the Team Lead Ladder if they will run plans across the team.
How do I help an architect or reviewer evaluate the design? Reviewer or Architect Ladder; pair with Appendix R — A Day in the Forge for three worked examples.
How do I find the canonical evidence and receipts? Chapter 1 — Evidence (A/B test results); the three vignettes in Appendix R reuse the same numbers in narrative form.
How do I show where Plan Forge sits in the GitHub stack? Appendix H — GitHub Stack Alignment; the per-component story is in Appendix I — Plan Forge on the GitHub Stack.