Skill catalog

Every skill, principle, and z-mode playbook in this library, read from the source files at build time. Skills are explicit-only unless marked model-invocable: invoke them by name, such as /how in Claude Code or $how in Codex. Playbooks load through z-mode.

Select an entry for examples, how it works, expected results, and requirements. Each reference includes a link to the underlying instructions.

Workflow skills

architect Sketch types, signatures, and module structure before code, then stay in the loop while implementation fills in. Use for /architect, 'architect this', 'design this', or non-trivial work where jumping to code would lock in the wrong shape.arena Spawn N parallel candidates at the same task, pick a base, graft the strongest parts of the losers into it. Use for /arena, 'arena this', 'throw it in the arena', or when one attempt at a non-trivial artifact would lock in the wrong shape.automate-me Draft or update a personal mode skill from stated conventions and explicitly scoped history.benchmark-checklist Vet a perf measurement (limiter, tuning, limits, errors, repeatability, relevance, and whether the work happened) before you report or act on it. Use when you run a benchmark or report a speedup or regression you measured.blast-radius Find what a change could break somewhere else before it ships, beyond the diff, and prove the one fact it's safe because of by running real code instead of writing it up. Use for 'blast radius of X', 'what could this break', or reviewing a small diff you don't trust.bro Restate the last message in plain human language, with no jargon.correct Find repeated agent mistakes in this repository and prevent recurrence through architecture, types, lint or CI, then behavioral tests, with docs last. Use for correct.create-formal-verification-skill Generate a reusable project-local skill for model checking or proving code-linked properties. Use explicitly for /create-formal-verification-skill, model-checking a race, or proving an invariant; UI and service journey checks belong to create-verification-skill.create-verification-skill Generate a project-local verification skill that drives your app the way a user does — any language, framework, or platform. Use for /create-verification-skill, "make a control skill for this repo", or when a project has no scripted way to prove UI/CLI/service behavior.figure-it-out Design an auditable playbook when no narrower one fits: a large migration, an ambitious multi-part change, or work a human reviews after stepping away. Scales rigor to the task, runs a hypothesis loop, and logs decisions via show-me-your-work. Use for /figure-it-out, 'figure it out', a large migration, or when no narrower playbook applies.herdr-workflow Use Herdr as the execution layer alongside existing zstack workflows when the user enables it; also handle explicitly requested Herdr agent and pane coordination.how Read-only explanations of how a named code subsystem works: entry points, callers, runtime flow, package ownership, and layering. Use for code walkthroughs and architecture questions. Use why for historical design rationale; do not select for implementation or debugging requests.in-your-own-words Restate in your own words what you think the user's goals are and what problem they're trying to solve. Use for in-your-own-words.interrogate Adversarial review of a requested diff or design; only findings that a rerun reproduction proves survive.ketchup Catch up on this chat since your last message in plain language, with current PR, CI, and agent status where available and enough context to decide what awaits you. Use for ketchup or a catch-up on this chat; use recall for context across sessions.maintain-formal-verification-skill Refresh a project's formal checks after covered code changes, distinguish verification drift from product bugs, and retain per-check coverage checkpoints. Use explicitly for /maintain-formal-verification-skill or auditing whether a model still describes the code.maintain-verification-skill Periodic pass that keeps a project's verification skill and feature map honest: parallel source readers per feature, one live session driving every feature, a scoped local set of proven corrections. Use for /maintain-verification-skill or "audit the verify skill".no-comments Review unnecessary comments and workarounds with a scoped comment-sicko report.python-best-practices Python best practices for writing or reviewing Python source, scripts, and packages with uv, Ruff, mypy, type hints, and Pydantic boundary validation.recall Reconstruct your recent working context from your own chat history, live state, and the shared record (user reports, prior fixes, incidents), then hand back a tight current-state brief. Use for 'recall my work on X', 'catch me up', 'what have I been working on', 'where did I leave off', before starting or resuming work.reflect Review the active session for durable lessons and propose or apply approved skill edits.setup-zstack model-invocableCheck zstack installation in Codex, Claude Code, or Hermes and configure requested native model settings.sharpen Rewrite a prompt with an expert approach, primary sources, specific methods and tools, and an independent check. Use for sharpen; does not execute the task.show-me-your-work Keep a reviewable decision trail for long-running or unattended work: a local TSV log with one row per decision (what, why, evidence, result). Commit only when authorized. Use for /show-me-your-work, autonomous or multi-phase runs, or work a human reviews after stepping away.statechart Check a TLA+ model of a system that moves between states, such as a product flow, a protocol, or a service, and draw it as a Mermaid statechart with the shortest path to each state that breaks a rule. Use explicitly for /statechart; render with Pele.swarm Fan out N parallel workers, drain them, and return one report. Use for /swarm, 'swarm this', or parallel coverage, races, gauntlets, and exploration.tdd Use only when the user explicitly asks for TDD, a failing test, or a regression test, OR when the bug has an obvious cheap local test target. Skip when the test path is unclear, expensive, integration-heavy, or not requested.teach Explain a body of work plainly so a person actually understands it. Runs the `how` and `why` skills and weaves what they find into one clear explanation. Use for 'teach me this', 'help me really understand X', 'explain this change or subsystem to me'.technical-writing Layered technical-writing standard: Diátaxis structure, Google developer style sentences, STE instruction rules, Global English syntax. Use for /technical-writing or when writing or reviewing docs, RFCs, readmes, PR descriptions, or commit messages.typescript-best-practices TypeScript best practices. Use when reading or editing any .ts or .tsx file.unslop Cut AI tells from requested writing while preserving meaning and intended tone.why Read-only research into historical code design decisions: why a subsystem works this way or why a particular approach was chosen. Uses relevant Git history and available supporting sources, separating documented facts from inference. Use how for runtime behavior; do not select for implementation or debugging requests.z-mode Evidence-driven engineering with scoped agents, portable principles, and task-sized verification. Use for z-mode; explicitly enabled Herdr execution applies across playbooks.

Principles

Principles are leaves that z-mode reads when they change a decision. See Use principles.

attack-the-premiseApply when two or more fixes sharing a premise fail the same gate. Write down and directly test that premise before another fix assumes it.boundary-disciplineApply when wiring validation, error handling, or framework adapters. Concentrate guards at system boundaries (CLI, config, network, external APIs); trust internal types and keep business logic in pure functions.build-the-leverApply to any non-trivial work, not just bulk work: edits, migrations, analyses, checks. Build the tool that does it or proves it (codemod, script, generator, or a skill your subagents follow) instead of working by hand. The tool is the artifact a reviewer can rerun.encode-lessons-in-structureApply when you catch yourself writing the same instruction a second time, or notice a recurring correction. Encode the rule as a lint, metadata flag, runtime check, or script instead of more text.exhaust-the-design-spaceApply when facing a novel UI interaction or architectural decision with no precedent in the codebase. Build 2-3 competing prototypes and compare side by side before committing.experience-firstApply when product, UX, or feature-scope tradeoffs come up. Choose user delight over implementation convenience; ship fewer polished features over more rough ones.explain-the-numberApply before you measure, and before you trust, report, or act on a number you measured: a speedup, a regression, a throughput, a latency, or an eval result. Find what limits it, and rule out that it measured something other than the work you think.fix-root-causesApply when debugging. Trace the first wrong decision or missing handling of a legitimate state, and fix it there; do not hide invalid states at the symptom.foundational-thinkingApply before writing logic: choosing core types and data structures, sequencing scaffold-vs-feature work, asking what concurrent actors share. Get the data structures right so downstream code becomes obvious.guard-the-context-windowApply when context is filling up: large outputs, long files, repeated reads, fan-out planning. Route bulk to subagents; keep summaries in the main thread, not raw payloads.laziness-protocolApply when refactoring, evaluating diff size, or tempted to add abstractions, layers, or signal threading. Bias toward deletion and the smallest change that solves the problem.make-operations-idempotentApply when designing commands, lifecycle steps, or processing loops that run amid crashes, restarts, and retries. Converge to the same end state regardless of partial prior runs.migrate-callers-then-delete-legacy-apisApply when introducing a new internal API while old callers still exist. Migrate callers and delete the old API in the same wave instead of preserving compatibility layers.minimize-reader-loadApply when reviewing or shaping code that's hard to trace. Count layers between question and answer, and hidden state in the reader's head; collapse one-caller wrappers and shrink mutable scope.model-the-domainApply when writing stateful logic, or when code branches a lot or repeats a shape assumption across files. Encode the domain in a structure instead of scattered conditionals.never-block-on-the-humanApply when tempted to ask 'should I do X?' on reversible work. Proceed, present the result, let the human course-correct after the fact; stay within authorized scope and reserve clarification for real missing decisions.outcome-oriented-executionApply during planned rewrites and migrations with explicit phase boundaries. Converge on the target architecture; don't preserve smooth intermediate states with throwaway compatibility code.prove-it-worksApply after completing a task, before declaring done. Verify against the real artifact (run the feature, read the actual value, inspect the diff), not a proxy, self-report, or 'it compiles.'redesign-from-first-principlesApply when integrating a new requirement into an existing design. Redesign as if the requirement had been a foundational assumption from day one, instead of bolting it on.separate-before-serializing-shared-stateApply when concurrent actors might write to the same file, branch, key, or state object. Eliminate the sharing first; serialize structurally only when one shared writer is a real invariant.sequence-verifiable-unitsApply to multi-step work (sweeps, migrations, runs of similar edits) and to how you stack commits and PRs. Break work into small units that each end in a verifiable state, check each before the next, and order delivery so the sequence proves itself to a reviewer.subtract-before-you-addApply when sequencing an addition, refactor, or rewrite. Remove dead code, redundant validators, and stub references first, then build on the simpler base.test-behavior-not-implementationApply when you write, change, or keep a test. Assert observable behavior so the test catches a broken contract and survives internal refactors.type-system-disciplineApply when designing types, reviewing a function signature, or writing code in any statically-typed language. Make illegal states unrepresentable, brand semantic primitives, parse external data at boundaries, refuse to lie to the compiler, exhaust variants, derive from authoritative schemas.

Playbooks

The z-mode router selects a playbook from the request.

Authoring a skillRead existing callers, metadata, resources, and invocation intent. Use installed native skill-authoring guidance when available; otherwise these steps suffice.Autonomous runState the checkable done predicate, scope, budget/stop condition, and available capabilities.Autopilot-fullExecute an independent queue with explicit done conditions. The name does not grant merge authority.Autopilot-stackRun Autopilot-full without merging. The operator reviews and lands.BabysitDeclare the mode from the request:Bug fixReproduce on the reported surface with the project's harness. If unreachable, record the attempted path and missing capability.EvalDefine the behavior under evaluation and 3-6 concrete criteria; hold the rubric out of candidate prompts.FeatureGround entry points, caller contracts, and existing patterns.HillclimbFix a metric, realistic workload, target, sampling method, correctness floor, and bounded attempt/time budget.InvestigationRead-only questions produce a cited explanation or recommendation.Multi-phase planPlanning produces a plan; implementation begins only when requested.Opening a PRRun only when publication is part of the user's request.OrchestrateYou run the program, never the code. Decide which agent does each piece of work, then check what each one produces. Route here for a standing multi-day program: many stacked PRs, many workers, and few human turns. Use autonomous-run for work that one agent could finish inside the session's budget.Pause safelyPause only on explicit pause/stop or when the session cannot continue.Perf issueCapture a repeatable baseline on the realistic workload that reproduces the complaint. Establish a measurable target for that workload before optimizing; if the request supplies none, agree on one with the user. Vet the baseline and each later number with benchmark-checklist.PrototypeName the decision this throwaway artifact should settle.RefactoringPin current behavior with existing checks or a focused characterization/equivalence check.Runtime forensicsThe deliverable is diagnosis, unless a fix was also requested.Session pickupRead the named handoff/session using the scoped evidence contract.ShippingOnly run merges when the user explicitly authorized landing.Trace forensicsIdentify the supplied artifact format and use an existing parser.Visual parityCapture or identify immutable baselines, states, viewport, fonts, and rendering environment before migration.Worktree cleanupResolve the bundled scripts/worktree_audit.py with uv run and pass the target repository explicitly. The default audit uses local Git only; --with-prs requests a read-only GitHub lookup.