Articles
When one tenant's CI noise blocks all the others
An agent in a clean session saw 153 unresolved CI alerts that belonged to another venture. Per-venture scoping plus a rollup view fixed it.
Per-minute session activity from JSONL: when heartbeats lie about idle time
Heartbeats only fire on /sos and /update. Per-minute activity from Claude Code JSONL gives sessions an honest activity timeline.
A Design System Sized for One Operator and a Team of Agents
Multiple ventures, one tokens package, three brownfield migrations, a CI gate, and the GA cut. What we built when the conventional playbook did not fit.
Rolling Out Six Claude Code Plugins Across a Fleet
Anthropic shipped a Claude Code plugin marketplace. We picked six, set precise triggers for each, and pushed them across the fleet in one session.
Migrating an Agent Fleet from Tarball Installs to a Private npm Registry
12 critical Dependabot alerts, one incomplete migration, and four PRs to finish the job. What the move from tarball URLs to GitHub Packages actually looks like.
Killing Skills on Purpose: A 32-to-24 Retirement Cascade
How a full-day audit cut a 32-skill catalog to 24 - and the operational tests that determined what earned retirement vs. what earned a stay.
Shipping a Cultural Directive Your Agents Read at Session Start
Captain's standing order kept getting forgotten. The fix: commit it to a file agents read at session start - and learn what text belongs there.
When a Behavioral Reflex Becomes a Slash Command
A directive repeated enough times got shipped as a skill. When that is the right call, and when memory or CLAUDE.md would have been the wrong tool.
Pursuing Partner Network Status as a Solo Operator with an AI Agent Team
A solo operator with an AI agent team applies to a partner program built for services firms. Transparent framing is the strategy, not the liability.