◼ Lab
Theworkbench.
The
workbench.
Experiments, prototypes, benchmarks, and field notes from SCTY’s applied research practice. We build things far enough to learn from them, keep what survives use, and fold the durable parts into client systems.
Systems
agency-cliAgents, handoffs, workflow state, and repeatable creative operations.↗phantom-cliBriefs into reviewed, publishable outputs.↗scty-mcpTool access for model-driven work.↗
shared stateA common record of work, decisions, and open loops.
Memory
wikiA living layer for sources, claims, voice, and decisions.
tokdownLong source material shaped for model use.↗pdf-cliReadable text and structure out of PDFs.↗civic-cliPolicy research with real citations.↗Loops
nightly loopDaily activity into briefs, reports, and next actions.
persona briefsContext and priorities that carry from week to week.
brandOSSignals, search, and model output in the same frame.↗field reportsObservations sharpened into public notes.↗Quality
tripwireFreshness checks, drift checks, and quiet failure detection.
autoresearchSmall experiments: measure, keep, revert, repeat.
critbenchCreative-process benchmark for judgment, coherence, voice, and refinement.↗givecare-benchEvals for high-stakes AI workflows.↗◼ Notes / Demos