morzecrew
- 30 skills
- 0 followers
- 8 hours ago last updated
- ▌ Drift To Gate · morzecrew bundleUse when the same rule is violated a second time, when a standing agreement lives only in prose or a memory file, when a gate rejects something genuinely legitimate, or when deciding whether an enforcement check still earns its keep. Not for a first violation.
- ▌ Error Taxonomy · morzecrew bundleUse when designing error handling for an API or module, deciding what to raise or what status an error maps to, reviewing raise sites and catch blocks, or sweeping a codebase where everything is a 500. Not for user-facing error copy.
- ▌ Escape Hatch Policy · morzecrewUse when someone asks for raw access, a passthrough field, a bypass flag, or a config override; when an API is tempted by a "pass anything" parameter; or when an existing hatch's usage is quietly growing. Not for feature flags or sanctioned lower layers.
- ▌ Failure Path Review · morzecrew bundleUse when writing or reviewing a consumer, worker, background loop, job runner, retry or backoff path, or shutdown handling; or after an incident involving stuck, lost, or duplicated work. Not for synchronous handlers, and not for diagnosing one live failure.
- ▌ Dependency Diligence · morzecrew bundleUse when about to add a package, library, or tool; when asked "should we use X" or "X vs Y"; or when reviewing a PR whose lockfile grows. Not for upgrading an already-adopted dependency, and not for choosing the platform itself.
- ▌ Determinism By Design · morzecrew bundleUse when code reads a clock, sleeps, generates random values or IDs, retries with jitter, or spawns concurrent tasks; when a test is flaky or timing-dependent; or when building replay or simulation. Not where the nondeterminism is the product, such as key generation.
- ▌ Authority Dissociation · morzecrew bundleUse when designing an approval, sign-off, waiver, or attestation; when a check passes work its own author produced; when an agent records a decision on someone else's behalf; or when one measurement is about to rule out a whole approach. Not for declared rule bypasses.
- ▌ Decide Before You Look · morzecrew bundleUse before any experiment, benchmark, A/B, spike, or evaluation; when a threshold is set or moved after results are visible; when a run is producing an inconvenient answer; or when someone asks for "3x better". Not for exploration that decides nothing.
- ▌ Fewer Tests More Proof · morzecrew bundleUse when asked to consolidate, dedupe, optimize, or clean up a test suite, to shrink its size or runtime without losing coverage, or to set up conformance testing across implementations of one contract. Not for writing missing tests, and not for deleting flaky ones.
- ▌ Python REST Docstrings · morzecrew bundleUse when writing, editing, or reviewing Python docstrings in a project that uses reST field lists — :param:/:returns:/:raises: and Sphinx cross-reference roles. Not for Google-style sections, not for NumPy style, and not for prose documentation.
- ▌ Ratchet What You Build · morzecrew bundleUse when finishing any guard, check, battery, or safe default; when a shipped check turns out to be opt-in or missing from CI; when a safe mechanism exists but the unsafe default still ships; or when asking why a protection did not fire. Not before the mechanism exists.
- ▌ Negative Result Taxonomy · morzecrew bundleUse when an experiment, candidate, prototype, spike, or benchmark fails and the next move is to start something else; when writing up a negative result; when triaging a backlog of dead attempts; or when a whole approach is about to be abandoned. Not when nothing was measured.
- ▌ Python Google Docstrings · morzecrew bundleUse when writing, editing, or reviewing Python docstrings in a project that uses Google style — Args:/Returns:/Raises:/Attributes: sections under Sphinx Napoleon. Not for reST field lists, not for NumPy style, and not for prose documentation.
- ▌ Composition Over Inheritance · morzecrew bundleUse when designing or refactoring a class hierarchy, deciding between extending and wrapping a class, or reviewing subclasses that override methods to disable them. Not for implementing an interface, or for a framework's documented subclass hook.
- ▌ Forze Skills · morzecrew bundleBuild backend services with Forze — specs, deps modules, handlers, governed aggregates, FastAPI routes, identity and tenancy, encryption, messaging, realtime, durable execution, analytics, and deterministic simulation testing. Use when writing or wiring an application that depends on forze and its integration packages.
- ▌ Rfc Writer · morzecrew bundleUse when asked to write an RFC, design proposal, design doc, technical spec, or architecture proposal; to record a decision and the alternatives it beat before building; to update an RFC's status after shipping; or to set up an rfcs/ directory. Not for user-facing documentation.
- ▌ Self Audit · morzecrew bundleUse when a branch, fix series, or document set of your own is finished and about to be merged or handed off, or when the user says "self-audit", "audit your work", or "check your own changes". Not for reviewing someone else's code, and not mid-flight.
- ▌ Altitude Docs · morzecrew bundleUse when writing, restructuring, polishing, or reviewing a documentation page — tutorial, quickstart, how-to, reference, explanation, landing page — or when deciding which archetype new content belongs in. Not for docstrings, RFCs, or changelogs.
- ▌ Readable Code · morzecrew bundleUse when naming or renaming anything, when code is deeply nested or arrow-shaped, when a comment restates what the code says, or when reviewing a diff for readability. Not when a name or a structure is dictated by an external contract or an ecosystem convention.
- ▌ Flag Dont Flip · morzecrew bundleUse when implementing against an RFC, spec, ADR, or design doc that carries graded decisions; when a decision needs to change mid-implementation; or when code has drifted from its spec. Not for authoring the design, and not for changes with no design document.
- ▌ Pr Review Loop · morzecrew bundleUse when a pull request has reviewer comments — AI or human — waiting to be addressed, or when a coverage gate is failing on one. Not for reviewing someone else's PR, not for pre-PR polishing, and never for merging.
- ▌ Distill The Rule · morzecrew bundleUse when a session ends in a genuine surprise, when a defect's shape will clearly recur, when a sabotage or test passes unexpectedly, after a postmortem, or when the same mistake class shows up twice. Not for one-off trivia, or anything the repo already records.
- ▌ Keep A Changelog · morzecrew bundleUse when asked to update CHANGELOG.md, add an entry, write release notes, or cut a version section; or when user-facing changes land that it should record. Not when the file is auto-generated in another format, and not for performing the release itself.
- ▌ Reading Isnt Proof · morzecrew bundleUse when one contract has two or more implementations — adapters, backends, clients, or a fake standing in for the real thing — and you are about to report that nothing tests something. Not for a single implementation, and not where a shared battery already exists.
- ▌ Reproduce Then Fix · morzecrew bundleUse when fixing any bug, flaky test, or incident; when a fix is proposed with no failing reproduction; when asked to just patch it; or when an error disappeared and nobody can say why. Not for triaging whether something is a bug at all.
- ▌ Gitmoji Conventional · morzecrew bundleUse whenever generating or suggesting a git commit message or Pull Request title, even when the user mentions neither gitmoji nor Conventional Commits. Not when the repository enforces a conflicting convention, and not when the user dictates the message verbatim.
- ▌ Less Code Same Behavior · morzecrew bundleUse when asked to deduplicate, DRY up, consolidate, or converge divergent code, or to shrink a codebase without changing behavior. Not when behavior is meant to change, and not for reviewing a single fresh diff.
- ▌ Python Testing Patterns · morzecrewImplement comprehensive testing strategies with pytest, fixtures, mocking, and test-driven development. Use when writing Python tests, setting up test suites, or implementing testing best practices.
- ▌ Measure Before Optimizing · morzecrewUse when about to optimize code or asked to make something faster, when reviewing an "X is faster than Y" claim, when writing or reading a benchmark, or when a design call carries performance consequences. Not when the problem is already measured and reproduced.
- ▌ Python Performance Optimization · morzecrewProfile and optimize Python code using cProfile, memory profilers, and performance best practices. Use when debugging slow Python code, optimizing bottlenecks, or improving application performance.