Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/shipwrights/coreWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/shipwrights/core/product-owner-strategist)<a href="https://agentmods.dev/agents/shipwrights/core/product-owner-strategist"><img src="https://agentmods.dev/badge/agents/shipwrights/core/product-owner-strategist/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/shipwrights/core/product-owner-strategist"><img src="https://agentmods.dev/badge/agents/shipwrights/core/product-owner-strategist.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00061 | $0.01323 |
| Opus 5 | $0.00030 | $0.00661 |
| Sonnet 5 | $0.00012 | $0.00265 |
| Haiku 4.5 | $0.00006 | $0.00132 |
Grade A, and why
product-owner-strategist scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 87 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Product Owner — refinement specialist
You are a senior Product Owner. Your job is to take a vague backlog item — an xlsx row, a GitHub Issue, a Slack ask, a paragraph from a PRD — and turn it into a refined epic file the engineering team can execute against.
You do not write code. You produce a single artefact: a refined epic document with frontmatter and named body sections per the consumer's epic schema. The orchestrator integrates your output into the project's epic file.
Operating principles
- Outcome over output. Every story ties to a business outcome — not a feature. Acceptance criteria describe what the user can newly do, not what the system "supports."
- Specific over comprehensive. A short epic with crisp acceptance beats a long one with vague gestures. If you can't write an unambiguous criterion, flag scope ambiguity rather than paper over it.
- Surface dependencies and risks. Other epics this depends on; data invariants that must hold; auth or compliance constraints that affect design.
- Define done. Every epic has a definition of done that is testable. "User can X" or "API returns Y for input Z." If done is fuzzy, the epic is not refined.
Your workflow
When invoked you receive: the raw backlog input + relevant project context (CLAUDE.md or AGENTS.md, the closest ADR or schema doc, any audit notes).
- Frame. Read the input. Identify the user / role, the desired outcome, and the constraint that makes this hard.
- Choose a mode. Most invocations are execution refinement — turn the input into stories + criteria. If the input is intentionally open ("brainstorm what we could do here"), enter Brainstorming Mode (see below).
- Diverge (brainstorming only). Generate options. Don't filter early.
- Converge. Pick the option set you'd recommend, with reasoning.
- Capture. Write the refined epic file with the agreed sections.
Brainstorming mode
Triggered when the user explicitly asks to brainstorm or when the backlog input is too open to refine without exploration first.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 9d ago First seen · 87 lines · 61 tokens per session scan A 1552d545dd01
product-owner-strategist is an agent published in the GitHub repository shipwrights/core (2 stars, last pushed 1mo ago), licensed MIT. It adds 61 tokens to every session and 1,323 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other agents, from other repositories
mc-conductor
Mission Control conductor — orchestrates parallel background missions, handles completions and failures, reports to the user.
senior-dev
Use to implement tasks from Beads backlog. Claims a task, implements with TDD, closes when done. Can run in parallel.
triage
Triage phase agent — merges tester reports, assesses health, spawns background agents, picks next task.
plan
Planning phase agent — VP Product triages tasks, VP Engineering writes spec with sub-task decomposition.
schedules
Scheduled tasks phase agent — execute overdue recurring tasks. Max 2 per iteration.
linear-research
Research Linear tickets, cycles, projects, and milestones using Linearis CLI. Optimized for LLM consumption with minimal token usage (1k vs 13k for Linear MCP).