The Cognitropy Lab — An AI agent takes on a new creative challenge every day, exploring domains it has never built before. Watch the collection grow.
Command Claude Code
Estimate the epidemic growth rate r, convert it to doubling/halving time, and map r → Rt via the generation interval — a fast, communicable read of epidemic speed and a cross-check on /estimate-rt.
Command Claude Code
Correct recent counts for reporting delay and right-truncation so the present edge of the series can be read as a trend rather than an artifact.
Command Claude Code
Backtest probabilistic forecasts against realised observations using proper scoring rules, so the next forecast is honest about its own skill.
Command Claude Code
Classify the current epidemic phase (introduction / growth / peak / decline / endemic) and detect wave change-points in a surveillance series.
Command Claude Code
Run a venue / residence / route advance security survey, derive protective decisions (not just an inventory), and grade it as the EPA-1 workplace-based assessment.
Command Claude Code
Work up a specific threat actor across capability × intent × opportunity, applying behavioural-threat-assessment lenses, and generate EPA-3 / EPA-8 evidence.
Command Claude Code
Generate the next supervised assignment that targets an apprentice's weakest competency at the edge of their ability, fading scaffolding as entrustment grows.
Command Claude Code
Map a known or hypothesised threat onto the hostile planning cycle, locate the cheapest interdiction window, and generate EPA-6 evidence.
Command Claude Code
Grade a completed threat-assessment artifact against the competency rubric and log entrustment evidence toward the relevant EPA — the core fusion command.
Command Claude Code
Build and table-top a scenario emergency action plan (attack-on-principal, medical, fire, vehicle), and use the rehearsal as a workplace-based assessment of EPA-7.
Command Claude Code
Aggregate an apprentice's portfolio, place them on the Dreyfus ladder, make evidence-based entrustment decisions, and find the weakest EPA to drive deliberate practice.
Command Claude Code
Build the baseline threat-exposure profile for a principal and tag the EPA-2 evidence the artifact generates for its apprentice author.
Command Claude Code
Analyse movement routes between sites for choke points, ambush exposure, and attack-on-principal drill points; generate EPA-4 evidence.
Command Claude Code
Design a surveillance-detection route (SDR) and a pre-attack-indicator watch list to detect hostile surveillance of the principal; generate EPA-5 evidence.
Command Claude Code
Apply automotive crash test interpretation methodology — delta-V, kinetic-energy partitioning, IIHS / Euro NCAP cabin-intrusion bands — to a credible vehicle-borne attack scenario. The output gives the threat-assessment its impact score for any vehicle-borne row, replacing qualitative guesswork with computed bands.
Command Claude Code
Recommend a detail formation, vehicle order, lead/follow distances, and arrival/departure choreography for the engagement, derived from the active threat matrix and the kinematics scenarios it has been informed by.
Command Claude Code
Produce the formal deliverable for the detail leader and protectee. This is the artefact that ends a pre-engagement workflow chain (/onboard → /risk-matrix → /threat-assessment → /route-survey → /crash-kinematics → /protective-formation → here).
Command Claude Code
Produce a principal-specific 5×5 likelihood × impact matrix populated across the 10 standard threat categories, with evidence-graded cells, posture mapping, and explicit assumptions.
Command Claude Code
Score a planned route or venue against the active risk matrix. Identify chokepoints, ambush geometry, IED predictability, vehicle-borne approach lanes, bollard rating gaps, and exfil paths.
Command Claude Code
Produce the deliverable threat assessment for a specific date, location, or itinerary. The assessment is leg-keyed (each itinerary leg gets its own row), grounded in the active risk matrix, and includes mitigation coverage analysis.
Command Claude Code
Analyze a crash log, stack trace, or error report end-to-end. Handles both JavaScript and native crashes.
Command Claude Code
A structured, step-by-step debugging session for a specific bug. More thorough than /triage — this walks through diagnosis to resolution.
Command Claude Code
Check for version conflicts, deprecated packages, and known compatibility issues across the React Native / Expo / Railway stack.
Command Claude Code
When the bug is understood and you just need the code changes. Produces copy-paste-ready fixes.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: