Live operations—Titanium Birch Labs

← Back to labs

What follows is a snapshot from the last site build: the same pipeline that generated this HTML read the repository, wrote a small JSON file under data/generated/, and the template below turned it into static markup. There is no server-side polling or external API—when the build runs again, the figures change.

This page is a small experiment in showing the site as a working system, not just a bundle of articles about one. It is not a polished dashboard, and it is not meant to suggest the operating model is finished. The point is simpler: if you are curious about how Titanium Birch works, you should be able to inspect some of that work directly.

Build time 2026-05-12 03:10 UTC

UTC, rendered into the page during the build.

Published articles 15

Current candidate-facing articles in the main sequence.

Recent changes shown 39

Commits on origin/main in the last 30 days (capped for length).

Open questions 8

Headings pulled from the open questions article.

Build freshness

What this build is based on

Generated at: 2026-05-12 03:10 UTC
Source commit: 77c295d
Recent change source: origin/main

This section shows the timestamp for the current site build, the commit used to build it, and the git reference used for the recent-changes list. There is no browser-side API call here.

Recent shipped changes

Latest changes on main

Each line is one commit on the branch named above, limited to the last 30 days and truncated if the list grows long. The headline is the subject line with common prefixes removed; the small tag next to it keeps the conventional-commit verb, merge marker, or ticket id when one is present.

2026-05-12 Bump @playwright/test in the npm-minor-patch group chore(deps-dev)
2026-05-12 Mermaid XSS hardening — strict defaults, no safeHTML (MAG-388) fix(security)
2026-05-12 Replace node:fs.globSync in SRI spec; pin Node 22 on deploy (MAG-387) fix(ci)
2026-05-11 SRI for Mermaid, Fathom exemption guardrail (MAG-385) feat(security)
2026-05-08 Lead Investment processes landing with PortfolioPenguin and DealDuck fix(MAG-382)
2026-05-08 Strip SVG blocks in extract-main-text before tag removal fix(MAG-381)
2026-05-07 Align confidentiality article with repo visibility key fact fix(MAG-379)
2026-05-06 Publish confidentiality article with inlined threat model and SOURCES Merge commit
2026-05-06 Add follow-up link for uneven DealDuck maturity (MAG-376) fix(content)
2026-05-05 Bump stylelint in the npm-minor-patch group chore(deps-dev)
2026-05-05 Bump stylelint in the npm-minor-patch group chore(deps-dev)
2026-05-04 Fix shell injection in audience-sim.yml that could exfiltrate CURSOR_API_KEY Merge commit
2026-05-04 Drop MagnetMagpie aside from agent-maturity; widen cross-section heuristic (MAG-371) fix(content)
2026-05-01 Align homepage Engineering signpost with Home charter (MAG-369) fix(content)
2026-04-30 Reshape skills-and-conventions git log to Trunk engineering; shift left on cross-section illustrative artefacts (MAG-368) fix(content)
2026-04-29 Point agent-maturity closing to open questions; require follow-up on acknowledged gaps (MAG-367) fix(content)
2026-04-28 Require concrete grounding in investment-processes articles (MAG-366) docs(skill)
2026-04-27 Bump stylelint in the npm-minor-patch group chore(deps-dev)
2026-04-27 MAG-365 audience sim grounding and repo naming fix(content)
2026-04-27 Name the repo when 'this repository' would be ambiguous (MAG-364) docs(skill)
2026-04-24 Address MAG-363 audience sim violations fix(content)
2026-04-23 Bump stylelint in the npm-minor-patch group chore(deps-dev)
2026-04-23 Freshness grid and 30-day activity context (MAG-360) feat(operations)
2026-04-23 Consolidate How we engineer openings (MAG-358) docs(content)
2026-04-23 Add depth anti-patterns to expectations register (MAG-359) docs(content)
2026-04-22 Resolve MAG-356 audience sim violations on site-automation and open-questions fix(content)
2026-04-21 Resolve MAG-355 audience sim violations on agent-maturity and code-review fix(content)
2026-04-17 Bump astral-sh/setup-uv in the actions-minor-patch group chore(deps)
2026-04-17 Lead with Labs brand and reading vs contributing (MAG-352) docs(readme)
2026-04-17 Set opus-4.7 as full model; fix site-automation audience-sim reference (MAG-351) chore(pm)
2026-04-16 Neutral meta description, sitemap, and robots.txt (MAG-350) fix(seo)
2026-04-16 Add 1200×630 Open Graph PNG for social previews (MAG-349) Merge commit
2026-04-15 Bump stylelint in the npm-minor-patch group chore(deps-dev)
2026-04-15 Qualify day-in-the-life claims for audience sim (MAG-345) fix(content)
2026-04-14 Resolve MAG-342 audience sim violations on home and articles fix(content)
2026-04-14 Update audience sim instructions and fix project name in skills (MAG-341) docs
2026-04-14 Update pyyaml requirement from >=6.0 to >=6.0.3 chore(deps)
2026-04-13 Bump stylelint in the npm-minor-patch group chore(deps-dev)
2026-04-13 Neutral phrasing for audience sim (MAG-339) fix(content)

Article recency

Section balance and freshness

Each row is one published article. Recency is the latest commit touching the markdown source that feeds that page (tracked paths under docs/ or content-examples/ when the build copies from there, otherwise the generated file under content/). The bucket counts whole days since that commit as of 2026-05-12.

≤7 days
≤30 days
≤90 days
>90 days

Investment processes

Investment decision support Investment processes≤7 days
Portfolio analytics Investment processes≤30 days

How we engineer

Where we sit on the agentic maturity curve How we engineer≤30 days
Skills and conventions How we engineer≤30 days
How a feature ships How we engineer≤30 days
How testing works How we engineer≤30 days
The operating model How we engineer≤30 days
The tech stack How we engineer≤30 days
How code gets reviewed How we engineer≤30 days

About this Labs site

How this site handles confidentiality About this Labs site≤7 days
Annotated artefacts from the build About this Labs site≤30 days
How this site builds itself About this Labs site≤30 days
The open questions About this Labs site≤30 days
The human side of early operations About this Labs site≤30 days
Day 1 live replay About this Labs site≤30 days

Article coverage

What the site currently covers

This list shows every article in the main reading sequence, pulled from the generated content files. It changes automatically when articles are added, removed, reordered, or rewritten.

Portfolio analytics

How PortfolioPenguin powers P&L attribution and risk factor analysis—surfacing what drives returns and where exposures sit.

The operating model

How a five-person firm runs engineering with AI agents as the primary contributors.

Investment decision support

How DealDuck supports IC memo reviews, deal triage, and VC portfolio framework modelling—structuring the preparation so humans can focus on the call.

The tech stack

Why an agent-driven engineering team chose each technology—and the trade-offs behind those choices.

How testing works

Sharp requirements, fast suites, and where CI sits—how testing works as an edit loop, not a separate department.

Where we sit on the agentic maturity curve

Dependability signals, rough edges, and what a mature loop looks like when checks are green but the output still misses the brief.

How code gets reviewed

The code review process—what machines handle, what the reviewer agent focuses on, and how comments are usually tagged by priority.

Skills and conventions

How institutional knowledge is encoded so agents can start with shared context, and how that guidance keeps evolving as the team learns.

How a feature ships

The six-phase delivery pipeline—from vague idea to production, run by agents.

How this site handles confidentiality

Threat-model data flow and asset list, one mitigation paired with the content expectations register, and the full docs/SOURCES.md lineage tables—inlined from the repo.

How this site builds itself

Delivery mechanics for MagnetMagpie: Linear, GitHub Actions, Cursor agents, CI, auto-merge, deploy, workflow catalogue, and agent skills—described here at a level visitors can follow.

Day 1 live replay

Timestamped replay of MagnetMagpie's first operating day (March 2026): merges, workflow failures, Slack-driven rule changes, and CI guardrails added in response.

The open questions

Seven design tensions still visible in the agent-operated model: measurement, review depth, skill maintenance, loop latency, quality gates, scale, and governance.

Annotated artefacts from the build

Four illustrative snippets—content-authoring skill, AGENTS.md fact table, git log, audience-sim log—with short commentary.

The human side of early operations

What the human role looked like during MagnetMagpie's first weeks—feedback loops, real failures, and the shift from fixing outputs to refining the system.

Open edges

Questions the site is still surfacing

These headings are pulled from the current open-questions article. They make the unresolved parts easier to scan from one place.

Want the fuller picture?

The articles explain the current model in more detail. For the firm beyond engineering practice, see titaniumbirch.com.