diff --git a/workshop/09-agentic-editing.md b/workshop/09-agentic-editing.md index f8230882..855b2b88 100644 --- a/workshop/09-agentic-editing.md +++ b/workshop/09-agentic-editing.md @@ -32,15 +32,15 @@ Then send: /agentic-workflows [your request here] ``` -The skill recognizes three core task types for day-to-day workflow maintenance: - -| Task type | When to use it | Example trigger phrase | -|-----------|---------------|------------------------| -| **Edit** | Improve the agent brief or [frontmatter](https://github.github.com/gh-aw/reference/frontmatter/) | "update the workflow to …" | -| **Debug** | Investigate unexpected output or a failed run | "debug the workflow — it ran but …" | -| **Optimize** | Reduce token usage or tighten [permissions](https://github.github.com/gh-aw/reference/permissions/) | "optimize the workflow to reduce [AI Credit](https://github.github.com/gh-aw/reference/cost-management/) cost" | - -If you are working locally or in a Codespace without a Copilot session, the terminal path in each section below shows the equivalent manual change. +The skill recognizes three core task types — **Edit**, **Debug**, and **Optimize** — and routes your request to the matching prompt. If you are working locally or in a Codespace without a Copilot session, the terminal path in each section below shows the equivalent manual change. + +> [!TIP] +>
+> Optional Side Quest: Want the full task-type table with example trigger phrases, plus practice scenarios for matching a request to the right task type? +> +> Work through [Side Quest: How the `agentic-workflows` Skill Dispatcher Works](side-quest-10-03-skill-dispatcher.md), then come back here. +> +>
## Start With One Concrete Observation diff --git a/workshop/README.md b/workshop/README.md index 83fe2472..0d14b9a6 100644 --- a/workshop/README.md +++ b/workshop/README.md @@ -107,6 +107,7 @@ A hands-on workshop that takes you from zero to a fully automated, AI-powered wo - [Project Future AI Credit Costs with `gh aw forecast`](side-quest-26-01-forecast-costs.md) — full walkthrough of `gh aw forecast`: reading P10/P50/P90 output, using `--period week` and `--days 7`, forecasting all workflows, and deriving a `max-daily-ai-credits` value from the P90 figure; branches from [Step 26](26-manage-costs-and-budgets.md). - [Skill Injection Strategies — Hint, Fusion, and Inline](side-quest-29-01-skill-injection-strategies.md) — decision table, code examples, and a practice exercise for the hint, fusion, and inline strategies for wiring a `SKILL.md` into a workflow prompt; branches from [Step 29](29-skills-and-domain-knowledge.md). - [Recognizing Common Agentic Workflow Failure Modes](side-quest-22-01-failure-modes.md) — worked examples of empty data, tool error, timeout, and prompt drift failures with a match-the-fix practice exercise; branches from [Step 22](22-error-handling-and-resilience.md). +- [How the `agentic-workflows` Skill Dispatcher Works](side-quest-10-03-skill-dispatcher.md) — full task-type table with trigger phrases, plus practice scenarios for matching a request to the Edit, Debug, or Optimize path; branches from [Step 9](09-agentic-editing.md). ## Getting Started diff --git a/workshop/side-quest-10-03-skill-dispatcher.md b/workshop/side-quest-10-03-skill-dispatcher.md new file mode 100644 index 00000000..124ec215 --- /dev/null +++ b/workshop/side-quest-10-03-skill-dispatcher.md @@ -0,0 +1,88 @@ + + +# Side Quest: How the `agentic-workflows` Skill Dispatcher Works + +> _One skill, three jobs — knowing which trigger phrase to use saves you from writing prompts from scratch every time._ + +## :dart: What You'll Do + +Learn how the `agentic-workflows` Copilot skill routes your plain-English request to the right editing, debugging, or optimizing prompt, and practice choosing the right trigger phrase for a given situation. + +## :clipboard: Before You Start + +- You have `.github/skills/agentic-workflows/` in your practice repository (created during [Step 7](07-your-first-workflow.md)). +- You can open Copilot CLI with `gh copilot` in your Codespace terminal. + +## Steps + +### Understand the dispatcher model + +The `agentic-workflows` skill acts as a dispatcher: when you describe a workflow task in plain English and mention the skill by name, it routes your request to the right editing, debugging, or optimizing prompt and makes changes directly in your repository. + +You invoke it in Copilot CLI in your Codespace terminal: + +```bash +gh copilot +``` + +Then send: + +```prompt +/agentic-workflows [your request here] +``` + +### Learn the three task types + +The skill recognizes three core task types for day-to-day workflow maintenance: + +| Task type | When to use it | Example trigger phrase | +|-----------|---------------|------------------------| +| **Edit** | Improve the agent brief or [frontmatter](https://github.github.com/gh-aw/reference/frontmatter/) | "update the workflow to …" | +| **Debug** | Investigate unexpected output or a failed run | "debug the workflow — it ran but …" | +| **Optimize** | Reduce token usage or tighten [permissions](https://github.github.com/gh-aw/reference/permissions/) | "optimize the workflow to reduce [AI Credit](https://github.github.com/gh-aw/reference/cost-management/) cost" | + +If you are working locally or in a Codespace without a Copilot session, you can still make each type of change manually — open the workflow file, make the targeted edit, then run `gh aw compile` before committing. + +### Practice matching a request to a task type + +For each scenario below, decide which task type (Edit, Debug, or Optimize) fits best before revealing the answer. + +**Scenario A:** Your workflow's daily comment is too generic and you want it to explain *why* something matters, not just *what* happened. + +
+Reveal answer + +**Edit.** You're improving the task brief's content, not fixing a failure or reducing cost. + +
+ +**Scenario B:** Your last run finished green, but no comment or issue appeared anywhere in the repository. + +
+Reveal answer + +**Debug.** Something went wrong between "the agent decided to write" and "the write actually happened" — that's an investigation, not a content edit. + +
+ +**Scenario C:** Your workflow works correctly, but you noticed it consumes more AI Credits than you'd like on a workflow that runs daily. + +
+Reveal answer + +**Optimize.** The behavior is correct; you're tuning for lower token cost. + +
+ +### Try writing your own trigger phrase + +Pick one real observation about your own workflow (from a recent run) and write a one-sentence trigger phrase using the task-type table above. Send it to the skill in Copilot CLI and review the diff it proposes before accepting. + +## :white_check_mark: Checkpoint + +- [ ] You can explain what the `agentic-workflows` skill dispatcher does in one sentence +- [ ] You can name the three task types and one trigger phrase for each +- [ ] You correctly matched all three practice scenarios to a task type +- [ ] You sent one real request to the skill and reviewed its proposed diff + +**Return to the main adventure:** [Refine, Test, and Improve Your Workflow](09-agentic-editing.md)