microsoft · olivomarco · Jun 28, 2026 · Jun 28, 2026
diff --git a/modules/ghaw/challenges/0-00-setup/COACH.md b/modules/ghaw/challenges/0-00-setup/COACH.md
@@ -14,7 +14,7 @@
 - Ask for a show-of-hands "green terminal" check after the first 10 minutes.
 - Pair any participant who is blocked with a working neighbor; don't let one blocker stall the group.
 - `gh aw` installs automatically via `postCreate.sh` in both Codespaces and local dev containers. If a participant sees "command not found", the postCreate script likely failed — have them rebuild the container.
-- If `gh aw run ... --dry-run` fails with an API key error, GitHub Copilot is the default zero-config engine. Confirm the participant's account has an active Copilot subscription. For Claude/OpenAI/Gemini, point them to `github.com/settings/codespaces` to add secrets.
+- If `gh aw trial ... --dry-run` fails with an API key error, GitHub Copilot is the default zero-config engine. Confirm the participant's account has an active Copilot subscription. For Claude/OpenAI/Gemini, point them to `github.com/settings/codespaces` to add secrets.
 
 ---
 
@@ -25,7 +25,7 @@
 | `gh auth status` fails | Not logged in | Run `gh auth login` again; choose HTTPS |
 | Codespace build times out | Network / quota | Retry once; fall back to local dev container |
 | `gh aw --version` not found | Container image stale or `postCreate.sh` failed | Rebuild container (`Ctrl+Shift+P → Dev Containers: Rebuild Container`); or run `curl -sL https://raw.githubusercontent.com/github/gh-aw/main/install-gh-aw.sh \| bash` manually |
-| `gh aw run modules/ghaw/resources/examples/hello-world.md --dry-run` fails | AI engine key missing | Confirm Copilot subscription is active; or set `ANTHROPIC_API_KEY` / `OPENAI_API_KEY` / `GEMINI_API_KEY` as Codespace secrets and rebuild |
+| `gh aw trial modules/ghaw/resources/examples/hello-world.md --dry-run` fails | AI engine key missing | Confirm Copilot subscription is active; or set `ANTHROPIC_API_KEY` / `OPENAI_API_KEY` / `GEMINI_API_KEY` as Codespace secrets and rebuild |
 | Repository not accessible | Not in the right repo | Confirm the Codespace or local clone is for `microsoft/frontier-agenticdevops-hackathon` |
 
 ---
@@ -36,7 +36,7 @@ Before releasing the group to Challenge 1-01, confirm per participant:
 
 - [ ] `gh auth status` exits 0 and shows the correct username
 - [ ] `gh aw --version` returns a version string
-- [ ] `gh aw run modules/ghaw/resources/examples/hello-world.md --dry-run` completes without errors
+- [ ] `gh aw trial modules/ghaw/resources/examples/hello-world.md --dry-run` completes without errors
 - [ ] Repository is open and writable (or readable with a documented fallback)
 
 ---

diff --git a/modules/ghaw/challenges/0-00-setup/README.md b/modules/ghaw/challenges/0-00-setup/README.md
@@ -74,7 +74,7 @@ gh auth status
 gh aw --version
 
 # 4. Dry-run smoke test
-gh aw run modules/ghaw/resources/examples/hello-world.md --dry-run
+gh aw trial modules/ghaw/resources/examples/hello-world.md --dry-run
 ```
 
 > All four commands must succeed before you move on. If any fail, see **Common Blockers** in the coach guide.
@@ -84,7 +84,7 @@ gh aw run modules/ghaw/resources/examples/hello-world.md --dry-run
 | Command | Expected output |
 |---|---|
 | `gh aw --version` | Returns a version string (e.g., `gh-aw version 1.x.x`) |
-| `gh aw run modules/ghaw/resources/examples/hello-world.md --dry-run` | Prints generated workflow YAML; exits 0 |
+| `gh aw trial modules/ghaw/resources/examples/hello-world.md --dry-run` | Previews the workflow without executing it; exits 0 |
 
 ---
 

diff --git a/modules/ghaw/challenges/0-00-setup/meta.yml b/modules/ghaw/challenges/0-00-setup/meta.yml
@@ -17,7 +17,7 @@ success_criteria:
   - "`gh auth status` exits 0 and shows your username"
   - "`gh --version` returns gh >= 2.x"
   - "`gh aw --version` returns a version string"
-  - "`gh aw run modules/ghaw/resources/examples/hello-world.md --dry-run` completes without errors"
+  - "`gh aw trial modules/ghaw/resources/examples/hello-world.md --dry-run` completes without errors"
   - "`ls modules/ghaw/resources/examples/hello-world.md` confirms the embedded example exists"
 tags:
   - setup

diff --git a/modules/ghaw/challenges/1-01-morning-briefing/COACH.md b/modules/ghaw/challenges/1-01-morning-briefing/COACH.md
@@ -125,8 +125,7 @@ safe-outputs:
   create-issue:
     title-prefix: '[Morning Briefing] '
 
-engines:
-  - copilot
+engine: copilot
 
 tools:
   github:
@@ -162,7 +161,7 @@ Keep the summary concise — 5-7 sentences total.
 - `tools: github: toolsets: [issues, pull_requests]` — agent can query data
 - `safe-outputs: create-issue:` with title prefix — consistent naming
 - Body: clear, structured instructions with specific counts to gather and format
-- `engines: [copilot]` — free, reliable default
+- `engine: copilot` — free, reliable default
 
 **Production-ready reference solution:** See `coaches/sample-solutions/track-1/1-01-morning-briefing.md`
 

diff --git a/modules/ghaw/challenges/1-02-safe-and-sound/COACH.md b/modules/ghaw/challenges/1-02-safe-and-sound/COACH.md
@@ -120,7 +120,7 @@ A coach can verify completion when the squad can show:
     create-issue:
       title-prefix: "[...]"
     noop:
-  engines: ...
+  engine: copilot
   ---
   ```
 - Remind: "2 or 4 spaces per level, consistent throughout."
@@ -146,8 +146,7 @@ safe-outputs:
     title-prefix: '[TODO Scan] '
   noop:
 
-engines:
-  - copilot
+engine: copilot
 
 tools:
   github:

diff --git a/modules/ghaw/challenges/1-03-the-watcher/COACH.md b/modules/ghaw/challenges/1-03-the-watcher/COACH.md
@@ -123,8 +123,7 @@ safe-outputs:
   add-comment:
   noop:
 
-engines:
-  - copilot
+engine: copilot
 
 tools:
   github:

diff --git a/modules/ghaw/challenges/1-04-label-maker/COACH.md b/modules/ghaw/challenges/1-04-label-maker/COACH.md
@@ -77,7 +77,7 @@ A coach can verify completion when the squad can show:
 **Root cause:** Safe-outputs config doesn't actually specify an allowlist, or allowlist is too broad.
 
 **Coach response:**
-- "Your safe-outputs should specify exact labels. For example: `safe-outputs: add-labels: labels: [bug, feature, question]`"
+- "Your safe-outputs should specify exact labels. For example: `safe-outputs: add-labels: allowed: [bug, feature, question]`"
 - Show them: "This means the agent can ONLY apply these three labels. It can't invent others."
 - Verify: "Check your frontmatter — is the allowlist enforced?"
 
@@ -123,16 +123,15 @@ permissions:
 
 safe-outputs:
   add-labels:
-    labels:
+    allowed:
       - bug
       - feature
       - question
       - documentation
       - help-wanted
   noop:
 
-engines:
-  - copilot
+engine: copilot
 ---
 
 # Label Maker
@@ -156,7 +155,7 @@ Explain your reasoning in a comment on the issue summarizing why you applied (or
 - `permissions: contents: read` — minimal
 - `safe-outputs: add-labels:` with explicit allowlist — constrained labels
 - `noop:` path — escape hatch for unclear cases
-- `engines: [copilot]` — reliable default
+- `engine: copilot` — reliable default
 - Body: clear heuristics with keywords and examples
 - Includes feedback comment so user understands the classification
 

diff --git a/modules/ghaw/challenges/2-01-triage-bot/COACH.md b/modules/ghaw/challenges/2-01-triage-bot/COACH.md
@@ -108,7 +108,7 @@ Agent: Use grep results as a hint; apply your judgment and post comment.
 **Coach response:**
 - "Did you set `on: issues: types: [opened]` in the frontmatter?"
 - Show them: `issue_comment` (for comments) vs `issues` (for issue events)
-- Have them recompile: `gh aw compile triage-bot.md`
+- Have them recompile: `gh aw compile triage-bot`
 - Then create a fresh test issue to trigger it
 
 ### Pitfall 2: Label Not Applied (But Comment Posted)
@@ -187,8 +187,7 @@ safe-outputs:
     allowlist: [bug, feature-request, documentation, question, duplicate]
   add-comment: {}
 
-engines:
-  - copilot
+engine: copilot
 ---
 
 # Issue Triage Bot
@@ -216,10 +215,10 @@ Keep the comment warm and welcoming—this is our first interaction with the con
 **Why this works:**
 - `on: issues: types: [opened]` — fires immediately when new issue created
 - `permissions: issues: read` — minimal, agent doesn't write directly
-- `safe-outputs: add-labels: allowlist:` — gates to 5 safe choices
+- `safe-outputs: add-labels: allowed:` — gates to 5 safe choices
 - `safe-outputs: add-comment:` — permission granted to post comment
 - Body: natural language, clear decision tree, emphasizes friendliness
-- `engines: [copilot]` — free, reliable for classification
+- `engine: copilot` — free, reliable for classification
 
 ---
 
@@ -230,7 +229,7 @@ Suggest this breakdown:
 | Phase | Time | What to Do |
 |-------|------|-----------|
 | Read & orient | 3 min | Read the challenge, ask questions |
-| Write frontmatter | 5 min | Set up `on:`, `permissions`, `safe-outputs`, `engines` |
+| Write frontmatter | 5 min | Set up `on:`, `permissions`, `safe-outputs`, `engine` |
 | Write body | 7 min | Draft classification logic + comment template |
 | Compile & test | 8 min | Compile, create test issue, watch Actions tab |
 | Debug & refine | 5 min | Fix any label/comment issues |
@@ -266,7 +265,7 @@ If a squad finishes in <20 minutes:
 If a squad is stuck:
 
 - [ ] Is the workflow file named `triage-bot.md` in `.github/workflows/`?
-- [ ] Did they run `gh aw compile triage-bot.md`?
+- [ ] Did they run `gh aw compile triage-bot`?
 - [ ] Does `.lock.yml` exist? (`ls .github/workflows/triage-bot.lock.yml`)
 - [ ] Is frontmatter valid YAML? (spaces, not tabs; proper indentation)
 - [ ] Is `on: issues:` spelled correctly? (not `issue:`)

diff --git a/modules/ghaw/challenges/2-01-triage-bot/README.md b/modules/ghaw/challenges/2-01-triage-bot/README.md
@@ -49,7 +49,7 @@ Create a gh-aw workflow named `triage-bot.md` in `.github/workflows/` that:
 ## Success Criteria
 
 - [ ] `.github/workflows/triage-bot.md` exists with valid gh-aw frontmatter
-- [ ] Frontmatter includes: `on: issues:`, `permissions`, `safe-outputs:`, `engines`
+- [ ] Frontmatter includes: `on: issues:`, `permissions`, `safe-outputs:`, `engine`
 - [ ] Safe-outputs has an allowlist limiting labels to: `bug`, `feature-request`, `documentation`, `question`, `duplicate`
 - [ ] `.github/workflows/triage-bot.lock.yml` compiles without errors
 - [ ] You manually test by creating a test issue in your repo

diff --git a/modules/ghaw/challenges/2-01-triage-bot/meta.yml b/modules/ghaw/challenges/2-01-triage-bot/meta.yml
@@ -13,7 +13,7 @@ prerequisite_capabilities:
 description: Build an event-driven workflow that categorises every new issue with an AI label and posts a friendly comment explaining the classification and next steps.
 success_criteria:
   - `.github/workflows/triage-bot.md` exists with valid gh-aw frontmatter
-  - Frontmatter includes: `on: issues:`, `permissions`, `safe-outputs:`, `engines`
+  - Frontmatter includes: `on: issues:`, `permissions`, `safe-outputs:`, `engine`
   - Safe-outputs has an allowlist limiting labels to: `bug`, `feature-request`, `documentation`, `question`, `duplicate`
   - `.github/workflows/triage-bot.lock.yml` compiles without errors
   - You manually test by creating a test issue in your repo

diff --git a/modules/ghaw/challenges/2-02-review-buddy/COACH.md b/modules/ghaw/challenges/2-02-review-buddy/COACH.md
@@ -192,8 +192,7 @@ permissions:
 safe-outputs:
   add-comment: {}
 
-engines:
-  - copilot
+engine: copilot
 ---
 
 # Review Buddy
@@ -220,7 +219,7 @@ Tone: helpful colleague, not gatekeeper.
 - `checkout: false` (implicit) — doesn't need code, saves time
 - `safe-outputs: add-comment: {}` — can post a single comment
 - Body: natural language, clear observations, emphasizes warmth
-- `engines: [copilot]` — free, reliable
+- `engine: copilot` — free, reliable
 
 ---
 
@@ -229,7 +228,7 @@ Tone: helpful colleague, not gatekeeper.
 | Phase | Time | What to Do |
 |-------|------|-----------|
 | Read & orient | 3 min | Read challenge, understand PR triggers |
-| Write frontmatter | 4 min | Set up `on: pull_request:`, `permissions`, `safe-outputs`, `engines` |
+| Write frontmatter | 4 min | Set up `on: pull_request:`, `permissions`, `safe-outputs`, `engine` |
 | Write body | 6 min | Draft observations + comment template |
 | Compile & test | 8 min | Compile, create test PR, watch Actions tab |
 | Refine tone | 6 min | Read comment, improve warmth/specificity |
@@ -265,7 +264,7 @@ If finished in <20 minutes:
 If a squad is stuck:
 
 - [ ] Is trigger `on: pull_request: types: [opened]`? (not `on: issues:`)
-- [ ] Did they compile? (`gh aw compile review-buddy.md`)
+- [ ] Did they compile? (`gh aw compile review-buddy`)
 - [ ] Does `.lock.yml` exist?
 - [ ] Did they create a fresh test PR *after* compiling?
 - [ ] Is there a workflow run in the Actions tab?

diff --git a/modules/ghaw/challenges/2-03-slash-and-burn/COACH.md b/modules/ghaw/challenges/2-03-slash-and-burn/COACH.md
@@ -210,8 +210,7 @@ permissions:
 safe-outputs:
   add-comment: {}
 
-engines:
-  - copilot
+engine: copilot
 ---
 
 # Slash Commands: /summarize
@@ -243,7 +242,7 @@ Be professional but friendly.
 - `checkout: false` (implicit) — doesn't need code
 - `safe-outputs: add-comment: {}` — can post summary
 - Body: natural language, clear instruction to check for `/summarize` first, then posts structured summary
-- `engines: [copilot]` — free, good at summarization
+- `engine: copilot` — free, good at summarization
 
 ---
 
@@ -289,7 +288,7 @@ If a squad is stuck:
 
 - [ ] Is trigger `on: issue_comment: types: [created]`?
 - [ ] Is `lock-for-agent: true` present?
-- [ ] Did they compile? (`gh aw compile slash-commands.md`)
+- [ ] Did they compile? (`gh aw compile slash-commands`)
 - [ ] Did they create a test issue + several comments + then post `/summarize`?
 - [ ] Is there a workflow run in the Actions tab?
 - [ ] Does the body check for `/summarize` before posting summary?

diff --git a/modules/ghaw/challenges/2-04-stale-patrol/COACH.md b/modules/ghaw/challenges/2-04-stale-patrol/COACH.md
@@ -175,8 +175,7 @@ safe-outputs:
   add-comment: {}
   update-issue: {}
 
-engines:
-  - copilot
+engine: copilot
 ---
 
 # Stale Patrol
@@ -204,7 +203,7 @@ Be respectful and helpful. These might be contributors' important work.
 - `permissions: issues: read` — can query issues
 - `safe-outputs: update-issue: {}` — can close (which is a form of update)
 - Body: clear 3-phase logic (query → warn → close), respectful tone
-- `engines: [copilot]` — good at date calculations and respectful copy
+- `engine: copilot` — good at date calculations and respectful copy
 
 ---
 
@@ -213,7 +212,7 @@ Be respectful and helpful. These might be contributors' important work.
 | Phase | Time | What to Do |
 |-------|------|-----------|
 | Read & orient | 3 min | Read challenge, understand scheduled triggers |
-| Write frontmatter | 4 min | Set up `on: schedule:`, `permissions`, `safe-outputs`, `engines` |
+| Write frontmatter | 4 min | Set up `on: schedule:`, `permissions`, `safe-outputs`, `engine` |
 | Write body | 10 min | Draft stale issue query + warn/close logic |
 | Compile & test | 9 min | Compile, manually trigger with `workflow_dispatch`, verify logic |
 | Refine | 3 min | Adjust dates/exemptions if needed |

diff --git a/modules/ghaw/challenges/2-05-welcome-wagon/COACH.md b/modules/ghaw/challenges/2-05-welcome-wagon/COACH.md
@@ -182,8 +182,7 @@ permissions:
 safe-outputs:
   add-comment: {}
 
-engines:
-  - copilot
+engine: copilot
 ---
 
 # Welcome Wagon
@@ -209,7 +208,7 @@ Keep the comment friendly and concise. New contributors should feel welcomed, no
 - `permissions: pull-requests: read` — can read PR context
 - `safe-outputs: add-comment: {}` — can post comment
 - Body: clear author_association check, warm tone, resource inclusion
-- `engines: [copilot]` — good at personalization and tone
+- `engine: copilot` — good at personalization and tone
 
 ---
 
@@ -218,7 +217,7 @@ Keep the comment friendly and concise. New contributors should feel welcomed, no
 | Phase | Time | What to Do |
 |-------|------|-----------|
 | Read & orient | 3 min | Read challenge, understand author_association |
-| Write frontmatter | 4 min | Set up `on: pull_request:`, `permissions`, `safe-outputs`, `engines` |
+| Write frontmatter | 4 min | Set up `on: pull_request:`, `permissions`, `safe-outputs`, `engine` |
 | Write body | 8 min | Draft author_association check + welcome message with resources |
 | Compile & test | 10 min | Compile, create test PR (or use alternate account), verify |
 | Refine tone | 3 min | Make comment warmer if needed |

diff --git a/modules/ghaw/challenges/2-06-mix-and-match/COACH.md b/modules/ghaw/challenges/2-06-mix-and-match/COACH.md
@@ -50,8 +50,7 @@ safe-outputs:
   create-discussion:
     category: "General"
 
-engines:
-  - copilot
+engine: copilot
 ---
 ```
 
@@ -153,7 +152,7 @@ A correct `2-06-mix-and-match.md`:
 
 ## How to Verify It's Working
 
-1. **Compile check:** `gh aw compile .github/workflows/2-06-mix-and-match.md` — no errors
+1. **Compile check:** `gh aw compile 2-06-mix-and-match` — no errors
 2. **Import check:** Inspect the compiled `.lock.yml` — the helper's content should appear merged into the agent context
 3. **Manual trigger:** Run via `workflow_dispatch` from the Actions tab
 4. **Discussion created:** Navigate to the repo's Discussions tab — a new post should appear in "General" within seconds of the workflow completing

diff --git a/modules/ghaw/challenges/2-06-mix-and-match/README.md b/modules/ghaw/challenges/2-06-mix-and-match/README.md
@@ -84,7 +84,7 @@ Make sure GitHub Discussions is enabled for your repository (Settings → Featur
 ### Step 4: Compile and validate
 
 ```bash
-gh aw compile .github/workflows/2-06-mix-and-match.md
+gh aw compile 2-06-mix-and-match
 ```
 
 No errors? You're ready to test. Trigger manually with `workflow_dispatch` to verify the Discussion appears.

diff --git a/modules/ghaw/challenges/3-01-the-relay/COACH.md b/modules/ghaw/challenges/3-01-the-relay/COACH.md
@@ -196,8 +196,7 @@ tools:
   repo-memory:
     file-glob: 'metrics/**/*.json'
 
-engines:
-  - copilot
+engine: copilot
 ---
 
 # Daily Metrics Collector
@@ -245,8 +244,7 @@ tools:
   repo-memory:
     file-glob: 'metrics/**/*.json'
 
-engines:
-  - copilot
+engine: copilot
 ---
 
 # Weekly Metrics Report

diff --git a/modules/ghaw/challenges/3-02-context-engine/COACH.md b/modules/ghaw/challenges/3-02-context-engine/COACH.md
@@ -190,8 +190,7 @@ tools:
   github:
     toolsets: [pull_requests, contents]
 
-engines:
-  - copilot
+engine: copilot
 
 checkout: false
 ---
-Original file line number
+Diff line change
@@ Expand Up / @@ -123,8 +123,7 @@ safe-outputs: @@
       add-comment:
       noop:
-    engines:
-      - copilot
+    engine: copilot
     tools:
       github:
@@ Expand Down @@