TrackCTL
Killed: Aug 24, 2026
Cause of Death
Computer-Use Substitution and No Validated Demand
Post-Mortem
TrackCTL explored config-as-code product analytics for PostHog: a CLI and SDK path that let coding agents define, deploy, and verify analytics without browser click-ops. The implementation and research were substantial, but the commercial wedge weakened as general computer-use automation became capable of operating the browser workflow directly. The remaining advantages—durable configuration, review, and drift detection—never produced TrackCTL-specific customer interviews, adoption, releases, or revenue evidence. Credential handling remained a real agent-workflow problem, but it is cross-cutting infrastructure rather than a reason to maintain a separate PostHog product; that problem now belongs in the EnvBank experiment. Continuing TrackCTL would duplicate automation capability and split attention without validated pull, so the standalone thesis was killed while its repository, archived implementation, and research were preserved.
Lessons Learned
A browserless agent workflow is not a durable product wedge when general computer-use systems can absorb the same task. Validate whether buyers require durable declarative state, review, and drift control—not merely automation—and keep cross-cutting credential infrastructure separate from a vertical wrapper.
What I'd Do Differently
Test the complete workflow with current computer-use agents before building a dedicated control plane, and obtain named customer evidence for configuration review or drift-management pain before investing beyond a thin prototype. Do not revive TrackCTL unless real PostHog teams demonstrate that declarative governance remains urgent after computer-use automation and EnvBank-style secret handling are available.