davidahmann · davidahmann · Aug 11, 2026 · Aug 11, 2026
@@ -0,0 +1,70 @@
+name: Field signal or learning resource
+description: Suggest one anonymized FDE lesson, failure pattern, useful resource, or missing capability topic.
+title: "[Field signal]: "
+labels:
+  - triage
+body:
+  - type: markdown
+    attributes:
+      value: |
+        Share one focused signal. Do not include customer names, private data, credentials, employer-confidential material, or vulnerability details. Use SECURITY.md for vulnerabilities.
+  - type: dropdown
+    id: signal_type
+    attributes:
+      label: Signal type
+      options:
+        - Anonymized delivery lesson
+        - Failure or anti-pattern
+        - Learning resource
+        - Missing capability or practice topic
+        - Terminology or role clarification
+    validations:
+      required: true
+  - type: textarea
+    id: signal
+    attributes:
+      label: What should practitioners learn?
+      description: State the lesson in your own words and keep it to one bounded claim.
+    validations:
+      required: true
+  - type: textarea
+    id: relevance
+    attributes:
+      label: Why it matters for FDE or applied-AI work
+      description: Explain which workflow, decision, delivery stage, failure, or operating responsibility it improves.
+    validations:
+      required: true
+  - type: dropdown
+    id: evidence_type
+    attributes:
+      label: Evidence type
+      options:
+        - Direct experience, anonymized
+        - Primary source or reproducible artifact
+        - Secondary source or practitioner commentary
+        - Question or hypothesis needing evidence
+    validations:
+      required: true
+  - type: input
+    id: source
+    attributes:
+      label: Source URL or public artifact
+      description: Optional for anonymized experience; use a stable public URL when the signal comes from a source.
+  - type: textarea
+    id: limits
+    attributes:
+      label: Limits and context
+      description: State where this may not transfer, what remains uncertain, and any vendor or environment dependency.
+    validations:
+      required: true
+  - type: checkboxes
+    id: checks
+    attributes:
+      label: Submission checks
+      options:
+        - label: I removed customer, employer-confidential, personal, credential, and security-sensitive information.
+          required: true
+        - label: I have the right to submit this text and have disclosed any affiliation or commercial interest in the description.
+          required: true
+        - label: I understand this is a research lead, not production evidence, authority, or an automatic repository recommendation.
+          required: true
@@ -6,7 +6,7 @@ This repository is a design and verification kit for production AI-enabled syste
 
 Treat an agent as one component option. For each consequential decision, first compare deterministic code, optimization, classical ML, retrieval, a foundation-model call, a bounded agent workflow, and human review as applicable. Select the smallest sufficient mechanism and preserve the authority, evidence, cost, fallback, and retirement rationale. `ARC-004`, `ARC-005`.
 
-Use [`README.md`](README.md) for the public entry door, the concise [`guide/README.md`](guide/README.md) for the human mental model, and this file as the working contract for repository navigation and changes.
+Use [`README.md`](README.md) for the public entry door, the concise [`guide/README.md`](guide/README.md) for the human mental model, [`guide/capability-roadmap.md`](guide/capability-roadmap.md) for role and practice orientation, and this file as the working contract for repository navigation and changes.
 
 ## Required orientation
 
@@ -25,7 +25,7 @@ Do not load the entire repository by default. Use one skill or task route, then
 
 | Path | Role | Treat it as |
 | --- | --- | --- |
-| [`guide/`](guide/README.md) | Gives a concise, linear explanation of the FDE method | Human orientation; narrative, not a normative production contract |
+| [`guide/`](guide/README.md) | Gives a concise explanation of the FDE method plus a role and practice roadmap | Human orientation; narrative, not a certification or normative production contract |
 | [`catalog.json`](catalog.json) | Lists governed artifacts, types, paths, and tags | Registry; update when a cataloged artifact is added, moved, or removed |
 | [`controls/`](controls/control-catalog.json) | Defines production requirements and release gates | Engineering policy normative within this guide |
 | [`schemas/`](schemas/README.md) | Defines valid structures for machine-readable artifacts | Structural source of truth |
@@ -66,6 +66,7 @@ Repository-local skills are instruction-only workflows. They grant no tool acces
 | Task | Read next | Expected result |
 | --- | --- | --- |
 | Learn or explain the method | [Concise FDE Guide](guide/README.md) → relevant Handbook or Engineering Kit link | Shared mental model without loading the complete repository |
+| Learn or assess FDE and AI-engineering capability | [Capability roadmap](guide/capability-roadmap.md) → one relevant practice mission → linked canonical artifacts | Role boundaries, capability gaps, and inspectable practice evidence without treating the roadmap as a hiring standard or production proof |
 | Lead an FDE or internal delivery engagement | [Concise FDE Guide](guide/README.md) when orientation is needed → [FDE playbooks](playbooks/README.md) → current lifecycle stage → required templates | Evidence-backed decisions from qualification through business-owned production operation |
 | Review an FDE or applied-AI portfolio | [Operate and Scale](playbooks/03-operate-and-scale.md) → [portfolio review](templates/fde-portfolio-review.md) → linked service reviews and field-learning records | Cohort-aware investment, continuation, productization, transfer, capacity, and exit decisions without overriding workflow gates |
 | Build shared applied-AI capability | [FDE and applied AI engineering synthesis](library/10-fde-and-production-agent-synthesis.md) → current lifecycle stage → relevant reusable artifact | A deliberate boundary between workflow-specific delivery, reusable product/platform capability, and sanitized field learning |
@@ -118,6 +119,7 @@ Do not begin with multi-agent topology or framework selection. First establish t
 - Executable tests provide regression evidence for the behavior they exercise; passing does not certify production readiness.
 - Blueprints and templates translate controls into reusable designs.
 - The concise Guide explains the shared mental model and routes readers to authoritative deeper artifacts.
+- The capability roadmap organizes learning and assessment over the same method; it is not a certification, production gate, or substitute for target evidence.
 - Library pages explain the reasoning; research records the evidence and its limits.
 
 If these disagree, do not silently choose one. Identify the conflict, preserve the safer behavior, and update every affected layer in the same change when feasible.

@@ -4,6 +4,10 @@ Notable changes to this project are recorded here.
 
 ## [Unreleased]
 
+- Add a compact FDE and AI engineer capability roadmap with role boundaries, eight capability domains, four evidence-backed practice missions, a five-part engagement starter pack, assessment prompts, and a concise glossary.
+- Publish the roadmap as a focused web route and integrate it into human, agent, catalog, contribution, and maintainer navigation without creating a parallel method.
+- Add a low-friction field-signal issue form for anonymized lessons, failure patterns, learning resources, and capability gaps while preserving confidentiality and evidence boundaries.
+
 ## [1.12.0] - 2026-08-10
 
 - Add a generated GitHub Pages guide with focused reader routes, accessible navigation, local search, canonical and social metadata, structured data, sitemap, crawler policy, and a machine-readable web index without creating a second content corpus.

@@ -55,6 +55,12 @@ Repository skills live under [`.agents/skills/`](.agents/skills/) and remain thi
 - be registered in [`catalog.json`](catalog.json) and covered by `npm run test:skills`;
 - update `README.md`, `guide/README.md`, `AGENTS.md`, and `llms.txt` when public navigation or the shared method changes.
 
+## Suggest a field signal or learning resource
+
+Use the focused [field-signal form](https://github.com/davidahmann/fde-guide/issues/new?template=field-signal.yml) for one anonymized delivery lesson, failure pattern, useful practitioner resource, or missing capability topic. This is the lowest-friction contribution path; you do not need to propose a control, schema, or implementation.
+
+Explain why the signal matters to FDE or applied-AI work, state whether it comes from direct experience or a source, and keep the claim no broader than the evidence. Do not include customer names, employer-confidential material, private data, credentials, or a disguised vendor pitch. A submitted signal is a research lead, not authority, target evidence, or an automatic roadmap addition.
+
 ## Pull requests
 
 Keep each pull request reviewable and scoped to one outcome. Explain:

@@ -10,20 +10,22 @@ An independent, open-source guide and engineering kit for forward deployed engin
 [![Latest release](https://img.shields.io/github/v/release/davidahmann/fde-guide)](https://github.com/davidahmann/fde-guide/releases/latest)
 [![License](https://img.shields.io/badge/license-Apache--2.0-blue.svg)](LICENSE)
 
-[Read on the web](https://davidahmann.github.io/fde-guide/) · [Read the concise Guide](guide/README.md) · [Use the Handbook](playbooks/README.md) · [Run the code](#see-it-working) · [Browse solutions](solutions/README.md) · [Use with an agent](#optional-use-it-with-a-coding-agent)
+[Read on the web](https://davidahmann.github.io/fde-guide/) · [Read the concise Guide](guide/README.md) · [Build FDE capability](guide/capability-roadmap.md) · [Use the Handbook](playbooks/README.md) · [Run the code](#see-it-working) · [Browse solutions](solutions/README.md) · [Use with an agent](#optional-use-it-with-a-coding-agent)
 
 ## Choose your depth
 
 This is one method at three levels. Start with only the depth your job requires.
 
 | Layer | Use it when | Start here |
 | --- | --- | --- |
-| **The Guide** | You want the mental model, core principles, and complete FDE delivery loop in about 20 minutes | [Read the concise Guide](guide/README.md) |
+| **The Guide** | You want the mental model, core principles, complete FDE delivery loop, or a practical path for building capability | [Read the concise Guide](guide/README.md) or follow the [FDE and AI engineer capability roadmap](guide/capability-roadmap.md) |
 | **The Handbook** | You are qualifying, designing, delivering, transferring, or operating a real workflow | [Follow the lifecycle playbooks](playbooks/README.md) and [human-readable library](library/00-start-here.md) |
 | **The Engineering Kit** | You need implementation artifacts, architecture, machine-readable contracts, release controls, executable examples, or tests | [Inspect the kit](#what-is-in-the-engineering-kit), [controlled-write system](examples/invoice-exception/README.md), and [hybrid system](examples/shipment-risk-triage/README.md) |
 
 The Guide explains the method. The Handbook supports judgment. The Engineering Kit makes claims, authority, behavior, and changes inspectable and testable. They are not separate frameworks.
 
+New to the role or assessing a team? The [capability roadmap](guide/capability-roadmap.md) compares adjacent responsibilities, organizes the work into eight capability domains, and provides four evidence-backed practice missions, a five-part starter pack, and a concise glossary. It is a learning route over this method—not a certification or separate framework.
+
 ## The core idea: engineer value before autonomy
 
 **Start with the work and the accepted outcome—not with a model or agent topology.**
@@ -60,6 +62,7 @@ The examples are in-memory teaching systems, not deployable products. Their test
 
 | You own | Start with | You should leave with |
 | --- | --- | --- |
+| Learning, hiring, or capability development | [FDE and AI engineer capability roadmap](guide/capability-roadmap.md) → one practice mission | Role boundaries, capability gaps, starter artifacts, and inspectable evidence without a stack-first curriculum |
 | Business value or use-case selection | [Concise Guide](guide/README.md) → [Discovery and Value](playbooks/01-discovery-and-value.md) | Bounded workflow, baseline, accepted outcome, verifier, value case, guardrails, and decision |
 | FDE or internal applied-AI delivery | [Handbook](playbooks/README.md) → current lifecycle stage | Evidence-backed path from observation through adoption, handoff, operation, and field learning |
 | AI engineering or software architecture | [Intelligence selection](library/12-software-architecture-and-intelligence-selection.md) → [blueprints](blueprints/README.md) | Smallest sufficient mechanism, coherent system boundary, contracts, failure behavior, and test plan |
@@ -178,7 +181,7 @@ git diff --check
 
 ## Contribute
 
-Start with [CONTRIBUTING.md](CONTRIBUTING.md). Use [Discussions](https://github.com/davidahmann/fde-guide/discussions) for design questions, [Issues](https://github.com/davidahmann/fde-guide/issues) for reproducible defects or evidence corrections, and the private channel in [SECURITY.md](SECURITY.md) for vulnerabilities.
+Start with [CONTRIBUTING.md](CONTRIBUTING.md). Use [Discussions](https://github.com/davidahmann/fde-guide/discussions) for design questions, the focused [field-signal form](https://github.com/davidahmann/fde-guide/issues/new?template=field-signal.yml) to suggest an anonymized lesson or learning resource, [Issues](https://github.com/davidahmann/fde-guide/issues) for reproducible defects or evidence corrections, and the private channel in [SECURITY.md](SECURITY.md) for vulnerabilities.
 
 Maintained by [David Ahmann](https://github.com/davidahmann) ([LinkedIn](https://www.linkedin.com/in/dahmann/)), a cloud, data, and AI platform leader with Field CTO experience. This is an independent project; no current or former employer endorsement is implied.
 

@@ -3,6 +3,7 @@
 | Need | Channel |
 | --- | --- |
 | Architecture or implementation question | [GitHub Discussions](https://github.com/davidahmann/fde-guide/discussions) |
+| Anonymized field lesson, failure pattern, learning resource, or capability gap | [Field signal](https://github.com/davidahmann/fde-guide/issues/new?template=field-signal.yml) |
 | Reproducible repository defect | [Bug report](https://github.com/davidahmann/fde-guide/issues/new?template=bug.yml) |
 | Control, pattern, or blueprint proposal | [Design proposal](https://github.com/davidahmann/fde-guide/issues/new?template=design-proposal.yml) |
 | Incorrect or stale evidence | [Evidence correction](https://github.com/davidahmann/fde-guide/issues/new?template=evidence-correction.yml) |

@@ -9,6 +9,12 @@
       "path": "guide/README.md",
       "tags": ["fde", "value", "architecture", "lifecycle"]
     },
+    {
+      "id": "guide.capability-roadmap",
+      "type": "standard",
+      "path": "guide/capability-roadmap.md",
+      "tags": ["fde", "ai-engineering", "capability", "learning"]
+    },
     {
       "id": "controls.production",
       "type": "control_catalog",

@@ -6,7 +6,7 @@ This document keeps the guide coherent as research, controls, templates, example
 
 | Layer | Authority | Change obligation |
 | --- | --- | --- |
-| `guide/` | Concise human mental model and route into deeper artifacts | Stay linear, readable, non-normative, consistent with the lifecycle, and free of duplicated contract detail |
+| `guide/` | Concise human mental model plus role and practice orientation | Keep the core method linear; keep the capability roadmap non-normative, evidence-oriented, and free of duplicated contract detail or certification claims |
 | `.agents/skills/` | Focused human- and agent-readable task routes | Keep triggers distinct, procedures thin, outputs explicit, metadata valid, and links bound to canonical artifacts |
 | `research/` | Dated evidence and caveats | Verify source, date, attribution, and claim boundary |
 | `controls/` | Normative project requirements | Link evidence and release gates; update affected verification |
@@ -66,7 +66,7 @@ This document keeps the guide coherent as research, controls, templates, example
 2. Run `npm ci --ignore-scripts`, `npm test`, and `git diff --check`; `npm test` includes solution, value-framework, skill-metadata, site-build, link, metadata, and catalog checks.
 3. Run spelling, action workflow, dependency, and secret scans used by the current project.
 4. Confirm all new governed artifacts are cataloged and every new source ID resolves.
-5. Proofread README, concise Guide, AGENTS, llms, generated site routes and descriptions, playbook routes, changelog, package/citation versions, and release links.
+5. Proofread README, concise Guide, capability roadmap, AGENTS, llms, generated site routes and descriptions, playbook routes, changelog, package/citation versions, and release links.
 6. When skill discovery or packaging changes, verify `npx skills add davidahmann/fde-guide --list` from a disposable environment; do not add this network-dependent smoke test to the deterministic CI gate.
 7. Use a scoped commit and draft pull request; do not bypass protected `main`.
 8. Require CI and review before merge; tag only after the release tree and metadata agree.

@@ -6,6 +6,8 @@ This is the concise human guide to forward-deployed engineering and internal app
 
 **Reading time:** about 20 minutes. This guide is narrative orientation, not a production standard or a substitute for the target organization's policy, security, architecture, or risk review.
 
+If you are learning the role, planning a development path, or assessing a team, use the companion [FDE and AI Engineer Capability Roadmap](capability-roadmap.md). It turns this method into role boundaries, capability evidence, four practice missions, a quick-start pack, and a glossary without creating a second methodology.
+
 ## 1. What an FDE is responsible for
 
 A forward-deployed engineer turns an ambiguous operating problem into a supported software service that produces a measurable outcome. The work crosses four responsibilities:
@@ -295,6 +297,7 @@ If a consequential answer is missing, stay in discovery. A model or framework ch
 
 Choose the depth that matches the work:
 
+- **Build capability:** use the [FDE and AI Engineer Capability Roadmap](capability-roadmap.md), complete one bounded mission, and keep the resulting limitations visible.
 - **Run the method:** use the [FDE Handbook](../playbooks/README.md) and follow the current lifecycle stage.
 - **Design a recurring solution:** start from the [business-flow portfolio](../solutions/README.md), then select only the relevant vertical profile and horizontal foundation.
 - **Build or review a system:** use the [Engineering Kit](../templates/README.md), [controls](../controls/control-catalog.json), [schemas](../schemas/README.md), [blueprints](../blueprints/README.md), [examples](../examples/invoice-exception/README.md), and [tests](../tests/).