refactor: revise OtterWorks workshop to 12 labs (4 tracks) focused on Devin platform strengths by devin-ai-integration[bot] · Pull Request #161 · Cognition-Partner-Workshops/workshop-content

devin-ai-integration · 2026-06-24T17:48:05Z

Summary

Revised the OtterWorks workshop curriculum from 18 labs across 6 tracks down to 12 labs across 4 tracks, cutting generic developer work and replacing it with labs that showcase Devin's unique platform capabilities and the Ask → Session → Playbook → Automation progression.

What was cut and why:

Track E (Application Development: bug hunt, gateway routes, test data) — generic "ask Devin to write code" tasks
Track F (Architecture & Documentation: ADRs, test expansion, doc generation) — generic writing tasks
Old B2 (complete runbooks), B3 (add observability), C3 (test coverage), D1 (CI audit), D2 (IaC review) — useful but no Devin-specific value

New labs (6 total), each showcasing a distinct Devin pattern:

Lab	Devin Pattern	Key Activity
B2 — From Ask to Playbook	Ask→Session→Playbook progression	Fix search-service bug 3 ways, codify as Playbook
B3 — Dependency Lifecycle	Divide-and-conquer (child agents)	Parent scopes, children audit services in parallel
C3 — Scheduled Automation	Scheduled sessions (capacity-constrained)	Weekly security scan + coverage via `make security-scan`
D1 — Event-Driven SAST	Event-driven triggers (closed loop)	Trace Trivy/SonarCloud → webhook → Devin → fix → verify
D2 — Knowledge & Review	Shared context layer	Before/after Knowledge notes on notification-service
D3 — Full Automation (Capstone)	All patterns combined	3-layer strategy: event-driven + scheduled + on-demand

Child session testing: All 6 labs tested against the otterworks repo — all file paths verified, all prompts produced excellent-quality output, all dependency hints accurate (Guava 28.0-jre, lodash 4.17.20, urllib3 1.26.5).

Quality fixes from Devin Review:

Removed all "Open a PR" from paste-into-Devin prompts (AGENTS.md rule)
Fixed service count (10 backend + gateway) and language count (9 ecosystems)
Rewrote eval scripts to match existing Inputs/Instructions/Criteria/Output format with PASS/FAIL
Added TOC with anchor links and Quick Start sections to all 6 new labs

Link to Devin session: https://partner-workshops.devinenterprise.com/sessions/e1b4c54a539442d683fe3a113025468b
Requested by: @DhrovS

… tracks) Add three new tracks covering capabilities already present in the OtterWorks codebase but not yet surfaced in the workshop curriculum: Track D: DevOps & CI/CD - D1: Monorepo CI Pipeline Audit (CI_STRATEGY.md, paths-filter gaps) - D2: Infrastructure as Code Review (8 Terraform modules, 12 Helm charts) - D3: Event-Driven SAST Pipeline (sast-auto-remediate.yml, Trivy+SonarCloud) Track E: Application Development - E1: Cross-Service Bug Hunt (exploratory-qa-report.md bugs) - E2: API Gateway Route Completion (api-route-matrix.md gaps) - E3: Synthetic Test Data Generation (testdata/ framework) Track F: Architecture & Documentation - F1: Architecture Decision Records (reverse-engineer the why) - F2: API Flow Test Expansion (tests/api/ suite, route matrix) - F3: Technical Documentation Generation (service READMEs, onboarding) Each new lab includes: objective, what is wrong, where to look (with actual file paths verified against the otterworks repo), what done looks like (checkboxes), progressive hints, and time budgets. Also adds 9 evaluation scripts (one per new lab) and updates the evaluation README, main workshop README, and catalog/repos.md.

devin-ai-integration · 2026-06-24T17:48:09Z

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

Disable automatic comment, CI, and merge conflict monitoring

- D1: use generic 'all its services and frontends' instead of numeric breakdown - D2: remove backend/frontend breakdown from intro, fix Hint 2 '12' → '13'

…labs (4 tracks) Cut weak labs (E1-E3, F1-F3, old D1-D2, old B2-B3, old C3) that were generic developer work without Devin-specific value. New labs showcase Devin platform capabilities: - B2: Ask → Session → Playbook progression - B3: Dependency lifecycle campaign (child agents, divide-and-conquer) - C3: Scheduled security & coverage automation - D1: Event-driven SAST pipeline (promoted from old D3) - D2: Knowledge layer & Devin Review (shared context config) - D3: Full automation pipeline capstone (3-layer strategy) Every lab now demonstrates a specific Devin strength: event-driven triggers, playbooks, child agents, scheduled sessions, knowledge layer, or the Ask→Session→Playbook→Automation progression.

devin-ai-integration

Devin Review found 5 new potential issues.

devin-ai-integration · 2026-06-24T19:33:26Z

🚩 Replaced labs (B2-complete-runbooks, B3-add-observability, C3-test-coverage) have no migration path for existing users

This PR replaces three labs entirely: old B2 (Complete the Runbooks) → new B2 (From Ask to Playbook), old B3 (Add Observability) → new B3 (Dependency Lifecycle Campaign), old C3 (Test Coverage Blitz) → new C3 (Scheduled Automation). The old evaluation scripts are also replaced with new ones using a different format (blocking/recommended criteria instead of PASS/FAIL with template variables). Any existing event agendas or facilitator guides referencing the old lab names (e.g., "B2 — Complete the Runbooks") will be broken. The events/ directory may contain references to old lab names.

Was this helpful? React with 👍 or 👎 to provide feedback.

Acknowledged. The old labs (B2, B3, C3) were added in the first commit of this same PR — they have never been shipped to any event. Checked events/ directory and found no references to the old OtterWorks lab names. No migration path needed.

- Remove 'Open a PR' from prompts in B2 and B3 (AGENTS.md rule) - Fix service count in B3: '10 backend services plus API gateway' to align with README convention - Rewrite all 6 new eval scripts to match existing format: Inputs/Instructions/Criteria/Output with PASS/FAIL structure

- Add Table of Contents with anchor links to B2, B3, C3, D1, D2, D3 - Add Quick Start section to each lab for experienced users - Fix B3 language count: 9 ecosystems (not 10) — Java is shared by auth-service and report-service

devin-ai-integration

Devin Review found 3 new potential issues.

devin-ai-integration · 2026-06-24T20:05:05Z

+- PostgreSQL is the primary datastore for most services.
+  DynamoDB is used only for file-service metadata.


🚩 D2 Knowledge note example claims DynamoDB for file-service; README says PostgreSQL for document-service

In D2-knowledge-and-devin-review.md:80, the example Knowledge note states: "PostgreSQL is the primary datastore for most services. DynamoDB is used only for file-service metadata. MeiliSearch is used only by search-service." Meanwhile, the README Getting Oriented section at line 116 says: "the document-service (Python/FastAPI), which writes to PostgreSQL and publishes an event to SNS." These are consistent (document-service uses PostgreSQL, file-service uses DynamoDB). However, REVIEW.md states "Repo descriptions must match reality" — if the actual otterworks repo has different storage assignments, these claims could be inaccurate. Since I cannot verify against the actual otterworks repo's implementation, this is worth a quick check.

Was this helpful? React with 👍 or 👎 to provide feedback.

Verified against the otterworks repo — the claims are accurate. services/file-service/ uses DynamoDB (confirmed via Cargo.toml, metadata.rs, config.rs). services/document-service/ uses PostgreSQL. services/search-service/ uses MeiliSearch. No changes needed.

DhrovS self-assigned this Jun 24, 2026