PO · strategy + requirements authority

PO · Product Owner

Strategic partner + requirements authority. Authors plans · grooms backlog · gates agent output against North Star. Authors governance documentation.

Pending uplift · M-2 + M-3
  • Agent file.claude/agents/po.md @ ae15ee20
  • Team membershipSpecialized direct-instance roster · Pattern A direct-instance per AGENT-OPERATING-MODEL.md §4 (PO main · PO1 memory-track per R37 §3-§4)
  • Direct-instance worktreeaiuni-uplift-po on s75-uplift-po
  • ToolsRead · Write · Edit · Bash · Grep · Glob · WebSearch
  • Catalog sectionAGENT-ROLES-AND-RESPONSIBILITIES-CATALOG §3.5
  • Existence statusexists

Summary

Jon's strategic partner and requirements authority. Conducts stakeholder interviews, grooms backlog, audits docs, triages issues, and gates all agent output against product requirements and the North Star.

Authors plans · scopes work · authors governance documentation · composes dispatch routing · reviews CC orchestrator output. Invoke for any strategic discussion, product direction question, or before finalizing Planner/Researcher/Validator output. PO scope discipline (S73 R23 codification — Correction 77): PO operates with execution-content scope intentionally bounded. Tool scope preserved at Read/Write/Edit/Bash/Grep/Glob/WebSearch — the boundary is which CONTENT PO writes (governance docs YES, platform code NO). PO DOES: Author plans · scope work · author governance docs · compose dispatch routing · author surface-back composition · open governance-doc PRs · review CC orchestrator output. PO DOES NOT: Write platform/product code · write test code.

Distinguishing characteristicsDifferent from code-pm because: PO authors content + dispatches (Read/Write/Edit/Bash/Grep/Glob/WebSearch), code-pm is read-only verifier. Different from Planner because: PO scope is product strategy + governance, Planner scope is curriculum content. Different from PMO/Chat because: PO is execution-content authority, PMO is read-only stenographer (per S75 C23 architectural-amnesia closure).

Research portfolio

Placeholder reasonPending M-2 L-8 corpus authoring (worktree `aiuni-uplift-po` clean from main HEAD `226c7036`; cascade L-4..L-8 lanes pending fire). Expected research portfolio per M-1 brief §PO class: product management methodology · stakeholder interview patterns · backlog grooming + WSJF + ICE prioritization · governance authoring + standing-orders discipline · scope-discipline + R12 quality-filter default. Sibling-cluster reference to Researcher Cluster 5 (methodology) + Cluster 6 (competitive analysis) for evidence-based decision-making.

Significant project contributionsS60 → Cumulative S58+

  • Cumulative S58+ · ongoing · Standing-orders + governance authoring
    Standing governance-authoring responsibility — CLAUDE.md amendments · AGENT-OPERATING-MODEL updates · RESPONSIBILITIES.md rules · BACKLOG entries · session-correction-log entries · agent-knowledge KB principles · skill files. R-numbered rules R1-R39 codified; SO #1-#31 standing orders maintained.
  • S73 PR #111 · 2026-05-02 · PO/CC execution boundary (R23 NEW)
    PO scope discipline codified — §1.5 PO DOES / PO DOES NOT contract. Closes Correction 77 root + sibling C79. ae15ee20
  • S60 · 2026-04-15 · PO Loop-Keeper Responsibilities (R13 NEW)
    Codified 7 explicit responsibilities — terminal-stall detection · checkpoint-timeout watch · CI-monitor sweep · drift-detection ping · re-release triggers · 3-bucket Terminal Roster format · cross-surface direction broadcast.

Learningsauthored + cross-cutting

Cross-cutting applied

  • R12 — quality-filter default (S60 origin)
    Quality-preserving option auto-defaults when decision surfaces quality-vs-convenience or literal-rule-vs-bypass. `User decision required` reserved for genuine strategic calls only.
  • R23 — PO/CC execution boundary
    PO authors content; CC executes git ops. PO does not write platform code or test code.
  • R13(g) — cross-surface direction broadcast
    When user delivers direction on one agent surface, PO proactively broadcasts to all active lanes whose committed state depends on prior direction. Same-turn, not deferred.

Skills + hooksused + constraining

Skills used (always-loaded)

Session-end Steps 1-10

session-end-execution

PO authors Steps 1-10 + sub-steps + emits Tier 2b standalone block. Verification owned by Code-PM agent (session-end-verification skill).

Pattern A/B/C decision matrix

multi-task-modes-and-delegation

Per-lane delegation routing across the three execution modes. Apply BEFORE every dispatch.

Evidence-based prioritization

prioritization

MoSCoW scoping · WSJF sequencing · ICE for experiments · outcome anchoring.

Skills task-match-loaded

11-phase ship workflow (SO #29)

deployment-signoff-proposal

Mandatory workflow for any product-surface ship. Forbidden patterns: direct-to-main pushes, mockup-as-validation-evidence.

User-approval gate (SO #30)

destructive-action-guardrails

Mandatory pre-action user-approval gate for ALL destructive or irreversible actions. 8 classes.

Hooks constraining

  • Tool-scope · agent file frontmatter (R23 codification)
    Read · Write · Edit · Bash · Grep · Glob · WebSearch — execution-content authority for governance docs. Boundary is content-scope (governance docs YES, platform code NO), not tool-scope.
  • PreToolUse:Bash gate
    Bash protection guard + hook router pre-validation before any Bash command execution.

Last updated · refresh details

  • Profile auto-regenerated2026-05-06T12:50:00Z
  • Refresh cadenceDaily midnight UTC · per-session-end · on-demand button (admin-tier optional)
  • agent file.claude/agents/po.md @ ae15ee20
  • corpusnot yet authored
  • catalogdocs/agent-knowledge/AGENT-ROLES-AND-RESPONSIBILITIES-CATALOG.md @ d7d1aee8+241affb8

Refresh strategy per HANDOFF v4 §12 — profile auto-regenerates from agent-file + corpus + executions + git log.

Cross-references from the relationship graph

Edges sourced from cross-refs/matrix.json · also browsable on the org chart.

Cross-references 4 skills · 2 hooks

Skills · always-loaded

  • Session-End-Execution
    Authors Steps 1-10 + sub-steps + emits Tier 2b standalone block.
  • Multi-Task-Modes-And-Delegation
    Pattern A/B/C decision matrix at HANDOFF_PACKET emit.
  • Prioritization
    Evidence-based prioritization for mixed backlogs.

Skills · task-match-loaded

  • Deployment-Signoff-Proposal
    11-phase workflow for any product-surface ship per SO #29.

Hooks constraining

  • tool-scope-frontmatter · hard
    Read · Write · Edit · Bash · Grep · Glob · WebSearch
  • claude-hook-router-pre · hard
    PreToolUse:Bash gate