Epic: Multi-Provider Agent Standardization Initiative

[ashleyshaw]

Epic Summary

Convert 16 ChatGPT/Codex agent exports in agents/ into standardised multi-provider agents compatible with Claude, GitHub Copilot, and OpenAI Codex, plus the repository infrastructure (schemas, hooks, instructions) that validates and governs them.

The Playwright Testing Agent is the Phase 1 pilot that establishes the pattern; the remaining 15 agents follow in Phase 2. Full planning package lives in .github/projects/active/agent-standards-initiative/.

Why: The exports are single-provider, inconsistently structured, and not validated. Standardising them gives one source of truth per agent, provider parity, and automated governance.

Linked Stories/Tasks

Phase 1 — Pilot + Infrastructure

• Playwright Testing Agent rewrite (pilot) — see child issue
• Schema framework (4 multi-provider schemas) — see child issue
• Validation hooks (4 agent hooks) — see child issue
• Instruction files + cookbook — see child issue

Phase 2 — Remaining 15 agents (one child issue each): ai-readiness-estimator, client-website-discovery-assistant, design-partner, harvest-analytical, linear-advisor, pagespeed, prd, prd-factory-planner, proposal-desk, tour-operator-config, website-content-strategist, website-scope-estimator, woo-config, wp-config, zendesk-support.

Child issue numbers are cross-linked in the comments once created.

Milestones & Timeline

• Phase 1 (pilot + infra): 12–18h — establishes the pattern
• Phase 2 (15 agents): 2–4h each, batched 2–3/week
• Phase 3 (governance consolidation): 4–8h

Acceptance Criteria

• All linked stories/tasks completed
• 16 agents converted to multi-provider format
• Schemas, hooks, and instructions merged and enforcing standards
• Documentation and changelog updated
• Tests/QA complete
• Milestone closed

Dependencies / Blockers

• Phase 2 agents depend on the Phase 1 pilot pattern being merged first.
• Infrastructure issues (schemas/hooks/instructions) should merge with or before the pilot.

Additional Context

• Planning docs: .github/projects/active/agent-standards-initiative/ (README, IMPLEMENTATION_SUMMARY, AGENT_STANDARDIZATION_AUDIT, PROMPT_1/2, batch prompts)
• Audit reports: .github/reports/audits/ and framework in .github/reports/frameworks/

---

Definition of Ready (DoR)

• Epic goal and scope defined
• Linked stories/tasks listed
• Milestones and timeline mapped
• Dependencies/blockers identified
• Estimate added
• Stakeholders/approvers listed (LightSpeed maintainers)

Definition of Done (DoD)

• All linked stories/tasks completed and closed
• Documentation/changelog updated
• QA and testing complete
• Milestone closed and release notes prepared

---

Related issues: (child issues linked in comments)

[linear-code]

LS-1635

[github-actions]

✅ Template check passed after update. Thanks for fixing the issue body.