Skill Claude CodeCodex
Route a Wildcat Labs request to the one canonical first-party or vendored skill that owns it, preserving the Promise Machine evidence boundary and every named sibling hand-off.
Skill Claude CodeCodex
Preserve heterogeneous lending-protocol captures by digest and expose a narrow, source-bound credit view for Tabularium and Probitas. Use when the user names Alexandria or asks to archive lending data for reproducible, address-scoped credit research. Raw release and registered Goldfinch and Clearpool derivation…
Skill Claude CodeCodex
Read and write the evidence statements that keep a release joined to the record behind it: an in-toto statement, optionally inside a DSSE envelope, with a predicate registry and gates that keep absence visible. Use when someone hands over an attestation and asks what it actually covers, when a release needs evidence a…
Skill Claude CodeCodex
Build, verify and evaluate evidence-backed protocol-agent releases: pin a document corpus by digest, prove citations as exact bytes, bind live values to a chain and block, separate source classes, grade recorded answers against an evaluation corpus and keep promotion as records. Use when the user names Berean, asks to…
Skill Claude CodeCodex
Enforce evidence-preserving structural output budgets on engineering prose. Apply automatically to chat answers and written drafts containing audit findings, security or diff review, gas analysis, invariant discussion, protocol analysis, or specification commentary, and on explicit $brevitas invocation. Govern volume…
Skill Claude CodeCodex
Optimise Solidity gas usage with an executable, fail-closed Foundry loop that measures savings, re-runs behaviour tests, checks storage layouts and method identifiers, and demands targeted differential or property evidence for state-sensitive unchecked arithmetic. Every candidate names a rule from a pinned corpus of…
Agent Codex
get in loser, we're building god.
Agent Codex
get in loser, we're building god.
Agent Codex
get in loser, we're building god.
Agent Codex
get in loser, we're building god.
Skill Claude CodeCodex
Work a failure you already have down to its cause, then guard it: stop the line, preserve the evidence, reproduce, localise, reduce to a minimal case, fix the cause rather than the symptom, and leave behind a test that fails without the fix. Use when a test fails, a build breaks, a fuzz campaign returns a…
Skill Claude CodeCodex
Decide what a step must emit so its behaviour is visible afterwards: the on-call questions first, then structured events, bounded metrics, correlation across the whole path, and alerts on what someone actually feels. Use when adding logging, metrics, tracing or alerting, when shipping anything that runs unattended, or…
Skill Claude CodeCodex
Run the one-shot delivery loop: study, runbook, then per-step implement/audit/prose/push until a working prototype exists. Use only when a Wildcat contributor explicitly asks to start, run, resume, or report a Hexaemeron or Fiat delivery, including /hexaemeron:fiat forms. Do not infer activation from a similar task.
Skill Claude CodeCodex
Generate Echidna/Medusa-compatible Solidity fuzz suites from Foundry or Hardhat projects. Trigger on "fizz", "generate fuzz suite", "build fuzz harness", "stateful fuzzing", "fuzzing harness", "property testing", and "invariant suite".
Agent Codex
Role: Perform manual protocol analysis when x-ray is unavailable, and produce a protocol-understanding.md file that downstream steps (4, 6, 7, 9) can read back as their protocol context.
Agent Codex
Role: Read the completed fuzzing suite outputs (coverage, corpus, campaign logs, properties, handlers) and produce the final report.md. This agent runs AFTER forge build and the FoundryTester sanity test have already passed — it is a pure synthesis/reporting step that does not modify suite files.
Skill Claude CodeCodex
Convert English-language properties in PROPERTIES.md (produced by the Fizz skill) into Solidity assertions inside the existing fuzz harness, then flip their checkboxes. Trigger on "fizz-convert", "convert properties", "implement properties from PROPERTIES.md", "convert PROPERTIES.md to Solidity".
Skill Claude CodeCodex
Reconcile an existing Fizz harness with a changed source tree. Detects added/removed/changed contract functions, quarantines stale properties, regenerates drifted handler stubs, and refreshes the snapshot. Trigger on "fizz-sync", "resync fuzzing", "sync fuzz harness", "refresh fuzzing properties", "fuzzing drift…
Skill Claude CodeCodex
Decide what gets written down and where it lives: the decision record behind a choice that would be expensive to reverse, the comment that explains why rather than what, the runbook an alert points at, and the README somebody starts from. Use when a decision is made, an interface changes, an alert needs somewhere to…
Skill Claude CodeCodex
Check prose that ships for banned AI writing tells, technical terms used without evidence, and repeated structural formulae that survive a wordlist. Use while drafting, editing, or reviewing public prose, or when the user names a term to ban. This is the diagnostic gate; Vulgate performs a content-preserving rewrite.
Skill Claude CodeCodex
Rank the held Next Fiat jobs across explicitly in-scope, non-mature skills, select the most worthwhile job out of 100, set one durable goal or loop, run that job through Fiat, then repeat until no eligible frontier remains. Use only when the user explicitly asks for Kronos, for a repeated ranked Fiat frontier loop, or…
Skill Claude CodeCodex
Change performance only against a recorded measurement: baseline first, one change at a time, re-measure the same way, then keep it or revert it. Use when something is slow, when a budget exists, when a change might have cost time, or when a profile is needed before touching code. Do not use it for Solidity gas, which…
Skill Claude CodeCodex
Harden the off-chain surface: the Python that harvests, indexes, replays and releases, the commands it shells out to, the URLs it fetches, the secrets it holds, the dependencies it pulls, and the model output it acts on. Use when a step accepts data from outside the process, runs a subprocess, fetches a URL, reads a…
Skill Claude CodeCodex
Hold a Fiat study and runbook to a content contract before any code is written: stated assumptions, testable success criteria, a chosen design with its trade named, the five discipline questions answered before any build, and steps that are discrete, green at both ends, sized for the audit loop and explicit about the…