Claude Skills by cognitive-fab

Turn Polygraph verification artifacts into clean, brand-consistent diagrams (SVG, optional PNG). polyviz is a DETERMINISTIC, artifact-derived renderer — same inputs produce byte-identical output and it makes no model call at render time. Use when the user wants to "visualize / diagram the verification", "draw the state machine / lifecycle", "render the counterexample / the bug the gate found", "show the fleet / compat / deploy-gate verdict", or "regenerate the figures for a report or post". I...

Author stateful code capture-ready BY CONSTRUCTION, in ANY language — so verification never needs an instrumentation retrofit. Apply whenever writing or substantially reshaping a state machine, workflow, reducer, store, or protocol handler in any session and any language (Python, Go, Rust, Java, TS outside polygen, …): shape the module to the eight CR requirements (one named step boundary, declared observable projection, observable rejections, a step-listener seam from the first commit, injec...

Write NEW stateful code that is verifiable from the moment it's written, instead of auditing code that already exists (that's the "polygraph" skill). Draft a contract from a feature description, author a SAM v2 strict-profile module against it (named intents/schemas/domains, keyed acceptors, reject(reason), sealed model, next-state (prime) acceptors), self-repair against reachable invariant violations before anything ships, and synthesize a demo/regression trace corpus. Use when the user want...

A polygraph for your state machine. Audit a stateful piece of code end-to-end: YOU (the agent) instrument a copy, build any test doubles needed to run it, and capture real execution traces, then derive a transition-function spec from its source with an LLM (default artifact: a SAM v2 strict-profile module — named intents/schemas/domains, keyed acceptors, observable reject(reason), sealed model, next-state (prime) acceptors), replay the traces against it, model-check it against invariants, and...

Elicit the invariants for a state machine — the plugin takes the lead. Harvest candidate invariants from the contract's own vocabulary (terminal states, typed fields, reject rules, effect kinds), contribute the frontier model's domain knowledge as candidates, pre-check every candidate against the machine so each question arrives with a HOLDS verdict or a concrete counterexample story, then drive a plugin-led interview that converges an append-only intent ledger into a confirmed invariants.mjs...

Check whether a state-machine version change is safe to ship against the live fleet, and produce the migration when the shape changed. Classifies the change into compatibility lanes (shape, vocabulary, intent, semantic, migration, composition), runs the mechanical gates those lanes require against fleet snapshots — setState round-trip, cross-version stimuli replay, migration validation, and the seeded model check ("can any live state be driven to an invariant violation under the new rules") —...

Build an LLM-era feature as a WORKFLOW, not an agentic loop — the composition skill that drives the whole toolset end-to-end. Design the stages along the data flow with the user (each failure either bends back to ONE named stage carrying its exact signal under a bounded budget, or escalates to a human), front-load the contract and the never-happens invariants, author the machine with polygen, declare effects (determinism at the edges, the model as one well-scoped handler), prove the anti-loop...

Analyse a system's architecture into a structured, evidence-carrying model, then render that model as a set of validated interactive diagrams and a matching markdown document. Use when asked to map, diagram, document or explain the architecture of a codebase or design; to produce architecture diagrams that stay honest about what exists versus what is only designed; or to keep an architecture document and its diagrams from disagreeing. Prefer this over drawing a diagram directly.