Tutorial 12 · Beginner
Turn a Proven Workflow into a Codex Skill
Standardize one founder workflow only after two reviewed runs, then create and test a focused local Codex skill that preserves evidence, failure, approval, memory, permission, and external-action boundaries.
You will build
A tested local learner-created Codex skill package with a stable workflow contract, known-case and fresh-case receipts, an owner and version note, and either an optional reviewed scheduled draft or a complete manual recurrence checklist
You will strengthen
Knowing when a workflow is stable enough to standardize, defining a narrow reusable contract, testing failure behavior, and keeping human judgment and permissions visible
Learning map
Tools and capabilities
Tools
Capabilities
Before you begin
What you need
- Codex in the ChatGPT desktop app, Codex CLI, or the Codex IDE extension
- The extracted Founder Operating Workspace
- Two previously reviewed runs of the same real workflow, or both separately linked Sample Venture raw fixture manifests with their comparison results held closed
Practice assets
Download what you need
Founder Operating Workspace starterThe portable workspace containing the local `.agents/skills/` authoring location, versioned founder workflow methods, and safe review boundaries.Download →Founder workflow run receiptCopy to `.agents/skills/weekly-founder-review/references/RUN-RECEIPT-TEMPLATE.md`. Record exact inputs, method and skill versions, output, checks, failures, founder review, approved memory changes, and external actions.Download →Reproducibility checklistCopy to `.agents/skills/weekly-founder-review/references/REPRODUCIBILITY-CHECKLIST.md`. Verify input identity, method, output traceability, repeated results, approval, and separate external-action authorization.Download →Sample Venture raw known-case manifestCopy to `.agents/skills/weekly-founder-review/tests/known-case/input-manifest.md`. This maps every raw known-case dependency while keeping the expected result outside the skill inputs.Download →Sample Venture raw fresh-case manifestCopy to `.agents/skills/weekly-founder-review/tests/fresh-case/input-manifest.md`. This maps every raw fresh-case dependency while keeping its expected result outside the skill inputs.Download →Sample Venture held known-case result SV-SKILL-RUN-001Keep closed during the known-case run. Open only after saving the actual output and receipt, then copy to `.agents/skills/weekly-founder-review/tests/known-case/expected/weekly-founder-review.md`. After that evaluation, it may also be copied to `.agents/skills/weekly-founder-review/tests/fresh-case/inputs/prior-review.md` as the fresh case's legitimate prior-review input.Download →Sample Venture held fresh-case result SV-SKILL-RUN-002Keep closed during the fresh-case run. Open only after saving the recovered actual output and receipt, then copy to `.agents/skills/weekly-founder-review/tests/fresh-case/expected/weekly-founder-review.md` for evaluation.Download →Weekly founder review methodCopy to `.agents/skills/weekly-founder-review/references/METHOD.md`. This is the versioned method used by both fictional raw fixtures.Download →Weekly founder review checklistCopy to `.agents/skills/weekly-founder-review/references/CHECKLIST.md`. This defines the required input, output, receipt, approval, memory, and external-action checks.Download →Weekly founder review output templateCopy to `.agents/skills/weekly-founder-review/references/OUTPUT.md`. This is the named internal draft structure used by both raw cases.Download →Sample Venture company contextCopy to `.agents/skills/weekly-founder-review/tests/known-case/inputs/context/company.md` and `.agents/skills/weekly-founder-review/tests/fresh-case/inputs/context/company.md`.Download →Sample Venture customer contextCopy to `.agents/skills/weekly-founder-review/tests/known-case/inputs/context/customers.md` and `.agents/skills/weekly-founder-review/tests/fresh-case/inputs/context/customers.md`.Download →Sample Venture offer contextCopy to `.agents/skills/weekly-founder-review/tests/known-case/inputs/context/offer.md` and `.agents/skills/weekly-founder-review/tests/fresh-case/inputs/context/offer.md`.Download →Sample Venture public-signal logCopy to `.agents/skills/weekly-founder-review/tests/known-case/inputs/evidence/source-log.md`.Download →Sample Venture interview SV-INT-001Copy to `.agents/skills/weekly-founder-review/tests/known-case/inputs/evidence/interviews/SV-INT-001.md`.Download →Sample Venture interview SV-INT-002Copy to `.agents/skills/weekly-founder-review/tests/known-case/inputs/evidence/interviews/SV-INT-002.md`.Download →Sample Venture interview SV-INT-003Copy to `.agents/skills/weekly-founder-review/tests/known-case/inputs/evidence/interviews/SV-INT-003.md`.Download →Sample Venture interview SV-INT-004Copy to `.agents/skills/weekly-founder-review/tests/known-case/inputs/evidence/interviews/SV-INT-004.md`.Download →Sample Venture interview SV-INT-005Copy to `.agents/skills/weekly-founder-review/tests/known-case/inputs/evidence/interviews/SV-INT-005.md`.Download →Sample Venture pricing notesCopy to `.agents/skills/weekly-founder-review/tests/fresh-case/inputs/pricing-notes.md`.Download →Sample Venture pricing workbookCopy the unchanged binary workbook to `.agents/skills/weekly-founder-review/tests/fresh-case/inputs/pricing-assumptions.xlsx`.Download →Sample Venture reviewed leadsCopy to `.agents/skills/weekly-founder-review/tests/fresh-case/inputs/leads.csv`.Download →Sample Venture guideThe fictional-data boundary and route map. Raw manifests and held comparison results remain separate training fixtures and never become real business evidence.Download →
Step-one preview
Choose a repeated workflow
Objective
Select one founder workflow that has produced a reviewed result at least twice and prove that it is ready for a standardization test.
Visible result
You can see one genuinely repeated workflow, two reviewed evidence records, a named owner, stable elements, unresolved differences, and a local-only standardization decision.
Completion check
I linked two reviewed runs of one workflow and confirmed that standardization is supported rather than assumed.