Brackenmere / dev log

Tag

Agents

10 posts tagged Agents.

G39 6 min read

The quest ledger, and five worlds finally joined up

Brackenmere had five worlds that no quest sent you to. I asked for the quest ledger to be redrawn as a map, sent it to two AI reviewers to play it in their heads, and the agents built the chapter that threads those worlds into the story: the Five Stones.

G38 6 min read

Three new worlds in one night

One evening I asked for three new maze worlds that were not rectangular grids, with movement and emotion, and stepped away. By morning a drowned abbey, a giant's kitchen and a clockwork hall were on main. Then I played them.

A08 5 min read

Three AI designers, one brief

When the game felt like an errand chain, I asked three AI designers the same neutral question and kept every other opinion, mine included, away from them until they had answered. How that ran, and what a month of building this way taught me.

A07 6 min read

The team takes shape: root, owners and one reviewer

My kids' playtest notes became four pieces of work, each with its own agent, and one agent above them merging the results. Over four days that settled into a team: a root that orchestrates, owners that build, one reviewer, and a single thing that stays mine.

A05 5 min read

The agents measure themselves

Three times in a fortnight I asked the agents to stop building and measure how they worked: a retrospective from the commit log, a pass through three weeks of their own transcripts, and an overnight round that timed every wait. Each time the numbers corrected something I was sure of.

A06 7 min read

Making the agents cheaper to run

In one week the agents' running costs became the bottleneck: byte limits that stalled good work, ceremony nobody read, and test suites that ran for hours. I asked where the time was going, and most of the answer was to do less.

A03 6 min read

Audits that become queues: simplifying what the agents built

Agents add code readily and almost never delete it. Three times in a fortnight I stopped feature work and had the agents audit what they had built, and the lesson was less about the audits than about what happened next: findings only got fixed once they became a queue of small, named jobs.

A02 6 min read

Tests that step time instead of waiting

In the cloud the game ran at under one frame a second, and a test suite that waited in real time fell apart. Stopping the clock fixed it; five days later, an honest bill for tests, reports and handoffs made the whole loop cheaper.

A01 6 min read

Handoffs, one worktree per feature and standing rules

The first week of Brackenmere ran as several AI agents at once, each in its own worktree, each starting from one short page that said what was true. A week later the work moved to the cloud, my personal rules moved into the repo with it, and one session became five threads in a day.