lkc-studio
UserClaude Code plugins by LKC. Rigorous engineering methods packaged for agents — proven techniques that go unused because they demand more patience than a person can sustain.
Categories
Indexed Skills (9)
bisect
This skill should be used when something broke and the cause is unknown — when the user says "it worked yesterday", "which commit broke this", "this started failing", "narrow down why", "find the minimal reproduction", "which config line causes this", "which dependency upgrade broke the build", "bisect this", or presents a large failing input, config, or test that needs to be reduced to its essential cause. Provides a unified method for locating a cause by binary search over commits, inputs, config, dependencies, or environment, plus working ddmin and predicate-validation tooling.
mutants
This skill should be used when the question is whether a test suite is actually any good — when the user says "are these tests any good", "what am I not testing", "coverage is 100% but I still shipped a bug", "improve the tests for this", "is this well tested", "find gaps in my tests", "mutation testing", or asks for tests to be written for existing code and the useful ones need identifying. Measures which bugs the suite would actually catch by mutating the source, then triages the survivors into real gaps and equivalent mutants.
backtrack
This skill should be used when a regex might be exploitable or slow — when the user says "is this regex safe", "ReDoS", "catastrophic backtracking", "the regex hangs", "one request pins the CPU", "check my regexes for denial of service", or is writing a pattern that validates user-supplied input. Statically flags backtracking-prone regexes, then proves the dangerous ones by feeding them a crafted input and measuring the blow-up.
cents
This skill should be used when code handles money — when the user says "is my money handling correct", "float rounding bug", "the totals are off by a cent", "should I use Decimal", "check my billing/pricing/payment code", "audit financial calculations", or is about to write code that computes prices, totals, tax, balances, or currency. Finds every place money is computed in floating point, where rounding error silently compounds.
gatekeeper
This skill should be used before installing a skill, plugin, or MCP server from someone else — when the user says "is this skill safe", "check this before I install it", "audit this skill", "vet this plugin", "should I trust this", "scan this for malicious code", or pastes a GitHub link to a skill they are about to add. Reports what a skill actually touches next to what it claims to do, so the gap is visible before it runs with your permissions.
properties
This skill should be used when tests need to be stronger rather than more numerous — when the user says "write property tests", "property-based testing", "use hypothesis", "what should I be testing here", "my tests pass but bugs still get through", "write better tests for this", "what's always true about this function", or when a mutation-testing survivor needs a test that kills a whole family of bugs rather than one example. Provides the property pattern catalog and a tool that measures whether a property actually catches anything.
strays
This skill should be used when things start failing for no clear reason — when the user says "everything is timing out", "the machine got slow", "all my commands are failing", "why is this hanging", "builds keep timing out", "the dev server is already running", "port already in use", or when several unrelated commands fail in a row during a long session. Also use before writing code that spawns subprocesses, so the spawn cannot leak. Finds processes that outlived whatever started them, and prevents creating them.
mutants
This skill should be used when the question is whether a test suite is actually any good — when the user says "are these tests any good", "what am I not testing", "coverage is 100% but I still shipped a bug", "improve the tests for this", "is this well tested", "find gaps in my tests", "mutation testing", or asks for tests to be written for existing code and the useful ones need identifying. Measures which bugs the suite would actually catch by mutating the source, then triages the survivors into real gaps and equivalent mutants.
bisect
This skill should be used when something broke and the cause is unknown — when the user says "it worked yesterday", "which commit broke this", "this started failing", "narrow down why", "find the minimal reproduction", "which config line causes this", "which dependency upgrade broke the build", "bisect this", or presents a large failing input, config, or test that needs to be reduced to its essential cause. Provides a unified method for locating a cause by binary search over commits, inputs, config, dependencies, or environment, plus working ddmin and predicate-validation tooling.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.