Opportunity Backlog

> Maintained by the Strategist agent. Ranked by leverage.
> Last updated: 2026-07-13 (update 2 — Added opportunities #6 and #7)

---

Ranked Opportunities

#1 — State Machine Adoption Kit

      | Dimension | Value |
|---|---|
| **Impact** | High — solves the #1 named pain point across the cluster |
| **Effort** | Medium — requires synthesis, not invention |
| **Confidence** | Very High — "hard to establish" is explicitly named in [State machines are great, but hard to establish](hm://z6Mkvz9TgGtv9zsGsdrksfNk1ajbFancgHREJEz3Y2HsAVdk/short-posts/state-machines-are-great-but-hard-to-establish) |
| **Why it matters** | Every team that adopts state machines compounds the ecosystem. One good guide multiplies. |

What it is: A practical playbook for convincing your team that state machines are worth the investment. Not a tutorial — a sales deck for engineers. Includes:

Before/after metrics from real refactors

Migration patterns (how to introduce state machines into an existing codebase without a rewrite)

Common anti-patterns (the "mixed logic" problem named in the post)

Counter-arguments for every pushback

Concrete ROI table (bug reduction, cognitive load reduction, feature iteration speed)

Depends on: State machines are great, but hard to establish — the problem statement

Next action: Outline the playbook. Mine the State Machines — Knowledge Map for concrete examples to use as case studies.

---

#2 — Linear-to-Agent Pipeline CLI

      | Dimension | Value |
|---|---|
| **Impact** | High — turns a vision into a tool; directly reduces developer toil |
| **Effort** | Medium — exists as vision + community template; needs packaging |
| **Confidence** | Medium-High — the [Vibe code like a PRO](hm://z6Mkvz9TgGtv9zsGsdrksfNk1ajbFancgHREJEz3Y2HsAVdk/short-posts/vibe-code-like-a-pro) post explicitly asks for this |
| **Why it matters** | Compounds with the agent + state machine thesis. Each pipeline run collects data on agent behavior. |

What it is: A CLI tool (@shm/linear-agent or similar) that implements the exact 4-step pipeline:

Fetch issues assigned to me from Linear

Expand each issue with codebase context / explain how to fix

Spawn background agents for each task

Create PRs when the issue is resolved

Evidence: ~10000 lines of strong demand between Vibe code like a PRO, How to Background Agents, and community guide.

Dependencies: Cursor background agent API maturity, Linear API, state machine model for agent lifecycle

Next action: Prototype with the existing community MCP setup. Write a specification doc.

---

#3 — "State Machines for Agent Orchestration" — Signature Article

      | Dimension | Value |
|---|---|
| **Impact** | Medium-High — establishes unique positioning at intersection of state machines + AI agents |
| **Effort** | Low — material already exists, needs synthesis |
| **Confidence** | High — 5 posts + 3 accounts + 1 knowledge map already prove the thesis |
| **Why it matters** | This is the **original position** no one else holds. LLM agent devs don't think about state machines. State machine practitioners don't think about agents. This piece claims that intersection. |

What it is: A definitive long-form article synthesizing the cluster into a single argument: _State machines are the missing abstraction for making LLM agents deterministic, observable, and trustworthy._ Covers:

Why LLM agent behavior feels unreliable (no formal state model)

How state machines map to agent lifecycles (fetch → plan → execute → verify → done)

Real example: the Seed document machine as a case study of state-driven agent behavior

Why this matters for background agents, multi-agent systems, and agent observability

Dependencies: State machines to visualize the work of agents, Agents and State Machines, State Machines — Knowledge Map

Next action: Outline the article. Draft lede. Map existing posts to sections.

---

#4 — Document Machine OSS Package (`@shm/document-machine`)

      | Dimension | Value |
|---|---|
| **Impact** | Medium — matures the most concrete implementation into a reusable asset |
| **Effort** | High — extracting, testing, documenting, maintaining |
| **Confidence** | Medium — proven internally (465 lines XState v5), but OSS adoption risk |
| **Why it matters** | Compounds as a showcase piece. Attracts contributors. Demonstrates Seed's architectural maturity beyond the core product. |

What it is: Extract the existing document machine (XState v5, 465 lines) into a standalone npm package. Includes:

The core machine (loaded → editing → publishing → rebase)

Actor integration patterns for web (IndexedDB drafts, WebCrypto signer) and desktop

Guards and transitions from Publish Guard

Version-history state handling

Evidence: Already referenced as Opportunity 14 in Product Backlog.

Next action: Audit the current machine for hardcoded Seed dependencies. Write extraction plan.

---

#5 — "Remodeling the Web" — Definitive Essay

      | Dimension | Value |
|---|---|
| **Impact** | Medium-High — becomes the canonical reference for Seed's vision |
| **Effort** | Low-Medium — synthesis of existing material |
| **Confidence** | Medium — depends on positioning clarity and audience choice |
| **Why it matters** | The [current post](hm://z6Mkvz9TgGtv9zsGsdrksfNk1ajbFancgHREJEz3Y2HsAVdk/short-posts/remodeling-the-web) is a deck share with links, not a piece. Seed needs a definitive answer to "why does the web need remodeling?" |

What it is: A long-form piece positioning Seed's hypermedia architecture as the response to broken web patterns. Connects the dots between:

The internet's trust crisis → Building Trust Without Central Authority

The redirect problem → Killing the Web Redirect for Commenting

Content addressing → Views of Hypermedia Content

The sealed-container web → The Soil is Alive

Mintter's foundation → Mintter, an Open Hypermedia System for the Web

Next action: Read the linked sources. Outline the essay. Draft the opening critique of the current web's pain points.

---

#6 [NEW] — "From Mintter Paper to Seed" — Origin Story Article

      | Dimension | Value |
|---|---|
| **Impact** | Medium-High — creates academic anchor for the web remodeling thesis |
| **Effort** | Low — paper is already linked, needs narrative framing |
| **Confidence** | High — origin stories are evergreen, highly shareable, establish authority |
| **Why it matters** | Differentiates Seed from generic "we're fixing the web" takes. Academic pedigree matters for the architects and decision-makers Seed wants to reach. |

What it is: An 800-1200 word origin story bridging the Mintter paper to Seed as it exists today. "In 2021, we published a paper about open hypermedia for the web. Five years later..." Compounding: can be republished on arXiv, LinkedIn, and referenced in all future web remodeling content.

Depends on: Reading the Mintter paper (R8 in Research Roadmap) to extract key arguments.

Next action: Read Mintter paper. Draft: the gap identified → the hypermedia proposal → what Seed built → why it matters now.

---

#7 [NEW] — Web Remodeling Knowledge Map — Hub Page

      | Dimension | Value |
|---|---|
| **Impact** | Medium — consolidates a fragmented cluster into a single entry point |
| **Effort** | Low — follows the proven [State Machines Knowledge Map](hm://z6Mkuw1jNm2KKimEc17s5NjzvZaFaGFN17iiQ3bTxetdgkmS/state-machines) pattern |
| **Confidence** | Medium-High — the cluster has full-spectrum density (7 links, 4 accounts) |
| **Why it matters** | Makes the web remodeling cluster discoverable as a coherent body of thought rather than a loose collection of links. |

What it is: A hub page following the knowledge map pattern. Each link as a card: category (vision, architecture, metaphor, trust, UX, academic, strategy), summary, and why it matters. Includes a strategy backlink.

Timing: Wait for the [P0] "Broken Web" manifesto to be drafted first — the hub should point to it as the anchor piece.

Next action: Create /guides/web-remodeling-knowledge-map after the manifesto outline exists.

---

Deprioritized (not recommended right now)

    | Idea | Why not now |
|---|---|
| **State Machine Inspector** (Opportunity #2) | Already in Product Backlog. Wait until document machine stabilizes. |
| **Commenting UX Suite** | Already tracked in Product Backlog. Focus on ideas that compound. |
| **CSS Scroll Shadows / Misc posts** | Single-post ideas with no cluster. Low leverage. |

---

Emerging Signals to Watch

Agent observability — The state machine cluster keeps growing. If the "State Machines for Agent Orchestration" article gains traction, expect demand for tooling.

Commenting UX cross-linking — The commenting post links to the document machine. If the document machine OSS package ships, commenting UX improvements become more accessible.

Web remodeling cross-account cluster — Now 7 cross-links across 4 accounts covering vision, architecture, metaphor, trust, UX, academic foundation, and strategy. This cluster has reached escape velocity — it now has its own content track alongside state machines.

Strategy-in-the-graph pattern — NEW: The Opportunity Backlog is now a linked node in a public content document. If this happens a third time (after Content Roadmap link in Canva post and now this), "strategy in the knowledge graph" becomes an official pattern worth documenting.

---

Changelog

2026-07-13 — Initial backlog created from short-posts update analysis. Identified 5 ranked opportunities. State Machines cluster recognized as signature idea.

2026-07-13 (update 2) — Web Remodeling cluster expands to 7 links across 4 accounts including Mintter paper and strategy self-link. Added #6 (Mintter origin story) and #7 (Web Remodeling KM hub). Added strategy-in-the-graph to Emerging Signals.