Strategic Proposal

Spawn 5 specialized expert sub-agents in parallel, each proposing roadmap items from their domain. An Expert AI Council drafts the roadmap, a Strategic Critic challenges it with adversarial backpressure, and the Council finalizes with revisions. The result is published as a pinned GitHub issue.

Core principle: SIGNAL OVER FEATURES. Items require evidence of user demand before entering "Build Now" phase. Infrastructure prerequisites are exempt. The Critic ensures the council isn't inflating signal or sandbagging complexity.

Variant: V2MOM / strategic operating model → wiki plan

Use this variant when the user asks for a council to define a V2MOM, operating model, strategic principles, or a plan to add strategic synthesis to the wiki. Do not force the full roadmap/GitHub-issue publishing flow unless the user explicitly asks for a roadmap update.

1. Gather current product truth from README/docs/context/wiki plus live external signal if available.
2. Spawn a small lens-diverse council (e.g. Product/Founder, Systems/Ops, Market/Docs) rather than the roadmap-specific five expert roles.
3. Run one adversarial Strategy Critic after the council draft. The critic must challenge overreach, weak measures, contradictions, and wiki-scope mistakes.
4. Final synthesis should separate:
   • Council decision: Vision, Values, Methods, Obstacles, Measures.
   • Wiki plan: exact target entry, draft frontmatter/body, verification, and rejected scope.
5. Prefer one bounded provisional wiki entry first. Extra positioning/docs-IA entries are premature unless they hold distinct durable facts.
6. Keep wiki output as synthesis, not council minutes. Raw/source material belongs under .agro/knowledge/raw/; the tracked entry stays within the wiki word cap and starts confidence: provisional.
7. Add explicit approval gates for contested strategic wording (e.g. tagline, key nouns, whether a constraint is too narrow) before implementing file changes.

Session example and final V2MOM synthesis: references/open-harness-v2mom-council.md.

Decision Flow

flowchart TD
    A["Guard: gh auth status"] --> B{Authenticated?}
    B -->|No| SKIP["Log: SKIP — gh not authenticated"]
    SKIP --> MEM_SKIP[Memory Protocol]
    MEM_SKIP --> Z_SKIP[HEARTBEAT_OK]

    B -->|Yes| C["Gather context: IDENTITY, schema, routes, issues"]
    C --> D["Compose Current State Briefing"]
    D --> E["Spawn 5 experts IN ONE MESSAGE (parallel)"]

    E --> E1["Expert: Product"]
    E --> E2["Expert: Docs"]
    E --> E3["Expert: Security"]
    E --> E4["Expert: Registry"]
    E --> E5["Expert: Agent Systems"]

    E1 & E2 & E3 & E4 & E5 --> F["Strategic Council DRAFT (opus)"]
    F --> CRITIC["Strategic Critic<br>Challenge signal, feasibility,<br>phase assignments, dependencies"]
    CRITIC --> F2["Strategic Council FINAL (opus)<br>Incorporate critique, revise or defend"]
    F2 --> G["Find/create pinned issue (label: roadmap)"]
    G --> H["Update pinned issue body"]
    H --> MEM_OP[Memory Protocol]
    MEM_OP --> Z_OP["Report: roadmap updated"]

Instructions

1. Guard: gh CLI authentication

gh auth status 2>&1

If this fails, log [strategic-proposal] SKIP: gh CLI not authenticated → Memory Protocol → HEARTBEAT_OK. Stop.

2. Gather context

Read the following to build the briefing:

• AGENTS.md — stack, mission, URLs
• .agro/cli/, scripts/, install/ — orchestrator entrypoints and provisioning surface
• docs/ — GitHub-readable core docs; rendered docs site source lives in mifunedev/agro-web
• Open issues: gh api "repos/mifunedev/agro/issues?state=open&per_page=50"
• Repo stats: gh api repos/mifunedev/agro --jq '{stars: .stargazers_count, forks: .forks_count}'

3. Compose the Current State Briefing

Assemble a structured markdown briefing to pass to ALL 5 experts:

## Current State Briefing

### Product Vision
1. Document Open Harness — the parent framework for AI agent sandboxes
2. Let users promote their forks — fork registry/showcase
3. End goal: curate Docker registries with monthly licensing — SaaS marketplace

### App State
- Routes: [list from step 2]
- Prisma models: [list or "none"]
- Auth: none
- API routes: none

### Infrastructure
- Docker Compose + opt-in PostgreSQL 16 overlay
- CI/CD: GitHub Actions (lint, format, type-check, build, test, E2E)
- Release: SemVer → GHCR Docker image
- Agent: 8 skills, 7 sub-agents, 4 heartbeats

### Community Signal
- Stars: [N], Forks: [N], Watchers: [N]
- Open issues: [N] (list titles + reaction counts)
- Recent fork activity: [list]

### Gaps
1. User accounts + auth (CRITICAL)
2. Fork registry data model (CRITICAL)
3. Docker registry integration (HIGH)
4. Subscription/licensing model (HIGH)
5. Open Harness documentation (HIGH)
6. Testing (MEDIUM — 2 tests total)
7. Observability (MEDIUM — no health endpoint)
8. Agent autonomy gap (MEDIUM — plans but no implementation)

4. Spawn 5 expert sub-agents in ONE message (parallel)

Launch 5 Agent tool calls in a single message for parallel execution:

Expert	Perspective
Product	Data models, APIs, features
Docs	Documentation, fork showcase UX
Security	Auth, headers, access control
Registry	Docker registry, licensing
Agent Systems	Agent autonomy, Ralph loop

Worker model and effort follow .agro/skills/delegate/SKILL.md: operator selections and exclusions bind, and the advisor selects and records unspecified settings per task.

Each expert is a prompt for a bounded provider-native worker, not a repository agent definition. Use subagent_type: general-purpose (or a read-only built-in when the expert only reads) and put the perspective, the Current State Briefing, and the required output format in the prompt itself. There is no .claude/agents/ file to read — this repository authors no project agents.

Experts operate independently — they do NOT see each other's proposals.

5. Strategic Council DRAFT

Launch a single Agent tool call for the council worker — a provider-native worker whose prompt carries the council role:

Pass the council:

• All 5 expert proposals
• The Current State Briefing
• Instruction to query actual signal data (repo stats, issue reactions, fork activity)
• Instruction to produce a DRAFT roadmap (the council's first pass — not final)

Save the council's draft output for the next step.

6. Strategic Critic review

Launch a single Agent tool call for the strategic critic — a provider-native worker whose prompt carries the adversarial role:

Pass the critic:

• The council's DRAFT roadmap
• The Current State Briefing
• Instruction to query actual signal data independently (verify, don't trust the council)
• Instruction to challenge every "Now" phase assignment, every signal claim, and every complexity estimate

The critic provides adversarial backpressure — its job is to find what's weak in the draft and force revision.

7. Strategic Council FINAL

Launch a second Agent tool call for the council worker, reusing the same council role prompt:

Pass the council:

• Its own DRAFT roadmap from step 5
• The critic's review from step 6
• Instruction: incorporate valid criticisms and revise, or explicitly defend against each challenge
• Every challenge from the critic MUST be addressed — either the item moves phase, the score changes, or the council explains why the critic is wrong
• The output is the FINAL roadmap — this is what gets published

The council's final output becomes the pinned issue body.

8. Find or create the pinned roadmap issue

Search for existing:

gh api "repos/mifunedev/agro/issues?state=open&labels=roadmap&per_page=10" \
  --jq '[.[] | select(.title == "Product Roadmap")] | first'

If none exists:

gh label create roadmap --repo mifunedev/agro \
  --description "Product roadmap tracking" --color "0075ca" 2>/dev/null || true

gh issue create --repo mifunedev/agro \
  --title "Product Roadmap" --label roadmap \
  --body "<council output>"

Then pin it: gh issue pin <NUMBER> --repo mifunedev/agro

If it already exists, update:

gh issue edit <NUMBER> --repo mifunedev/agro --body "<council output>"

9. Report

• HEARTBEAT_OK (if skipped)
• Full report: pinned issue # + top 3 "Now" items + signal summary

Reference

Key Resources

Resource	Where it lives
Expert roles (Product, Docs, Security, Registry, Agent Systems)	Prompts written inline in step 4 of this skill
Strategic Council role	Prompt written inline in steps 5 and 7 of this skill
Strategic Critic role	Prompt written inline in step 6 of this skill
Worker type for every role above	A provider built-in (general-purpose, or a read-only built-in) — no repository agent file backs any of them
Worker boundary policy	/delegate — When a worker is justified