Skip to content

feat: add workshop step 28 — Orchestrate Multiple Agentic Workflows - #2414

Merged
pelikhan merged 3 commits into
mainfrom
copilot/training-plan-add-research-backed-training-node
Aug 2, 2026
Merged

feat: add workshop step 28 — Orchestrate Multiple Agentic Workflows#2414
pelikhan merged 3 commits into
mainfrom
copilot/training-plan-add-research-backed-training-node

Conversation

Copilot AI commented Aug 1, 2026

Copy link
Copy Markdown
Contributor

The workshop taught building individual workflows (sub-agents, memory, MCP, evals, cost controls) but had no step on composing them into a pipeline. The dispatch-workflow safe-output makes cross-workflow orchestration possible but appeared nowhere in the curriculum.

Changes

  • workshop/28-orchestrate-workflows.md — New advanced step covering the orchestrator pattern: read repository signals, route to a specialist via dispatch-workflow, enforce max: 1 to keep dispatch predictable. Includes a decision-table design exercise, both AI-agent prompt and terminal paths, and compile/run/verify loop.

  • workshop/README.md — Step 28 added to Part 2 curriculum table.

  • workshop/27-evaluate-workflow-quality.md — Footer changed from "return to hub" to Next: Orchestrate Multiple Agentic Workflows to preserve linear curriculum flow.

  • .github/skills/micro-environment-simulator/workshop-student-journey.js — Added step 27 and step 28 to STEP_FILE_ALIASES, STEP_IDS, and buildTransitions(). Step 27 was already missing from the simulator, causing the mapping assertion test to fail; this also fixes that pre-existing failure.

Key frontmatter pattern introduced

safe-outputs:
  dispatch-workflow:
    workflows:
      - daily-status
      - pr-reviewer
    max: 1

- Create workshop/28-orchestrate-workflows.md covering dispatch-workflow
  safe-output orchestration pattern
- Update workshop/README.md curriculum table with step 28
- Update workshop/27-evaluate-workflow-quality.md to add Next: link
- Update simulator workshop-student-journey.js to add step 27 and 28
  mappings (step 27 was missing, causing simulator test failure)

Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
Copilot AI changed the title [WIP] Add research-backed training node for orchestrating workflows feat: add workshop step 28 — Orchestrate Multiple Agentic Workflows Aug 1, 2026
Copilot AI requested a review from pelikhan August 1, 2026 00:59
@pelikhan
pelikhan marked this pull request as ready for review August 1, 2026 02:00

@github-actions github-actions Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Warning

Firewall blocked 1 domain

The following domain was blocked by the firewall during workflow execution:

  • awmgmcpg

To allow these domains, add them to the network.allowed list in your workflow frontmatter:

network:
  allowed:
    - defaults
    - "awmgmcpg"

See Network Configuration for more information.

Generated by ✅ AW Accuracy PR Reviewer for #2414 · 36.8 AIC · ⌖ 4.82 AIC · ⊞ 4.3K

Comment thread workshop/28-orchestrate-workflows.md Outdated
```

> [!TIP]
> Use `gh aw compile --watch` while you iterate on the brief so the lock file stays in sync without manual recompiles.

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Inaccurate flag: gh aw compile --watch does not exist.

The gh aw compile command supports --strict, --validate, --fail-fast, --purge, and --approve, but has no --watch flag. Running this command as written will fail with an unknown flag error.

Suggested fix: Remove the tip, or replace with a plain recompile instruction:

gh aw compile repo-orchestrator

Comment thread workshop/28-orchestrate-workflows.md Outdated
2. Check whether a GitHub issue with the title prefix "Daily Repository Status" was created today.

Based on what you find:
- If stale open PRs exist, dispatch the `daily-status` workflow.

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Routing logic is reversed compared to the decision table above.

The decision table (lines 51-52) correctly maps:

  • Stale open PRs → dispatch the PR reviewer
  • No status issue today → dispatch the daily-status reporter

But this template brief reverses the mapping:

  • Line 95: If stale open PRs exist, dispatch the daily-status workflow ← should be pr-reviewer
  • Line 96: If no status issue exists today, dispatch the pr-reviewer workflow ← should be daily-status

The same reversal appears in the AI prompt block above (lines 62-63).

Suggested fix: Swap the target workflow names to match the decision table:

- If stale open PRs exist, dispatch the `pr-reviewer` workflow.
- If no status issue exists today, dispatch the `daily-status` workflow.

@pelikhan

pelikhan commented Aug 1, 2026

Copy link
Copy Markdown
Contributor

@copilot apply reviews from github action actor

- Swap dispatch targets to match decision table: stale PRs → pr-reviewer,
  no status issue → daily-status (was reversed in both prompt block and
  terminal template)
- Remove tip referencing gh aw compile --watch which does not exist

Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>

Copilot AI commented Aug 1, 2026

Copy link
Copy Markdown
Contributor Author

@copilot apply reviews from github action actor

Applied both reviews from the AW Accuracy PR Reviewer in commit fix: correct reversed routing logic and remove nonexistent --watch flag:

  1. Reversed routing logic — swapped dispatch targets in both the AI prompt block and the terminal template so they match the decision table: stale open PRs → pr-reviewer, no status issue today → daily-status.
  2. Nonexistent --watch flag — removed the tip referencing gh aw compile --watch.

@pelikhan
pelikhan merged commit 5689668 into main Aug 2, 2026
@pelikhan
pelikhan deleted the copilot/training-plan-add-research-backed-training-node branch August 2, 2026 10:10
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[training-plan] Add research-backed training node: Orchestrate Multiple Agentic Workflows

2 participants