odin-code-advanced

Use when non-trivial code needs designing before implementation, or an existing codebase needs its architecture improved: surveying for shallow modules, locating a seam, deepening a structure, hardening with guardrails, or making code testable. Also handles surveying for shallow modules, locating seams, deepening structures, hardening with guardrails, and making code testable. Not for architecture diagrams, deploy, remote, credential, or irreversible changes.

Use when asked to validate a web app, CLI, API, or generated artifact against a source-blind behavior contract. Produces a structured pass/fail/blocked/out-of-scope report with anti-cheat probes and redacted evidence. Not for source or remote-system changes.

Use when asked to determine what a change could break before it ships. Returns confirmed risks, cleared items, and the cheapest pre-merge test that catches the real bug. Not for remote, credential, publish, deploy, or irreversible changes.

Use when bloated code needs clean re-derivation, or the user says "this module is bloated" or "break it and rebuild". Classifies old behavior as essential or residue, cuts residue, and leaves the verifier green. Not for untracked data or changes without VCS rollback.

Use when a candidate change must be produced without touching the working tree; an ephemeral worktree runs the declared command and returns a binary-safe patch plus its exit code. Not for remote, credential, publish, deploy, or irreversible changes.

Use when a requested sweep monitors CI failures over a bounded attempt window. Returns each root cause reproduced or classified non-actionable with any minimal verified patch as a proposal. Not for classifying one failure without patching — use classify-ci-failure.

Use when asked to run /clean-clean-cut to cut accumulated records and residue under an explicit destructive gate with version-control recovery. Don't use for untracked targets, changes without a version-control rollback, or git branch/worktree cleanup — use git-cleanup.

Use when a user wants to identify the true sources of complexity qualitatively before counting metrics. Returns a ranked complexity-source report with root causes before any metric is counted. Read-only; no source or remote mutation.

Use when asked to reproduce, profile, or verify CLI/TUI behavior. Produces a deterministic transcript or profile proof with session cleanup. Not for CLI design advice — use cli-for-agents.

Use when asked to verify or reproduce browser or Electron UI behavior by capturing before and after evidence with no leftover processes. Not for remote, credential, publish, deploy, or irreversible changes.

Use when asked to remove old code, migrate consumers, or decide whether to maintain or sunset a system. Produces a proven replacement, migrated consumers, removed obsolete tracked code, and rollback evidence. Not for untracked data or changes without VCS rollback.

Use when asked to design a module interface, seam, or testable boundary. Produce compared alternative interface designs using explicit deep-module vocabulary and deliberately placed seams. Not for UI direction picking — use design. No source or remote-system changes.

Use when code mixes units, fixed-point precisions, scaling factors, rates, prices, shares, or conversions and needs dimensional consistency validated. Annotates every in-scope file with unit comments, derives a units vocabulary, and reports confirmed and refuted mismatches. Not for type-level unit modeling — use type-driven.

Use when reconciling an API endpoint's generated OpenAPI schema and declared response types with its actual runtime response. Adds or fixes schema decorators, reuses canonical types, migrates legacy path definitions, and validates the spec locally. Not for general API documentation: use docs-and-adrs.

Use when asked to prove coverage, find missing cases, or enumerate a state, decision, requirement, or behavior space. Enumerates every cell as covered, gap, or deferred with an executed check per cell, and emits a coverage manifest with a tally. Not for round-based questioning — use askme; not for testing one property — use property-test-authoring.

Use when asked to run a performance campaign against a measured floor: rebuild hot paths from their floor, grade cold paths, and land each target as an atomic commit with a proven win. Not for hypothesis-only analysis without mutation — use fastopt.

Use when a user wants to optimize suspected hot paths without waiting for benchmarks: label every cost observation as a hypothesis and emit benchmark designs, with no mutation or speed claim. Not for measured-floor rebuild campaigns — use extremely-optimize.

Use when a user wants to optimize estimated hot paths and complexity-neutral cold paths while refusing complexity theater. Emits hypotheses and benchmark designs. Not for measured-floor rebuilds — use extremely-optimize; not for hypothesis-only — use fastopt.

Use when asked for multi-commit stack edits, rebases, fixups, or stacked-PR publishing with branchless git idioms and a publish gate. Not for plain-git workflows in repos without branchless; the skill stays inert there. Human-only.

Use when a repository decision needs an interview plus durable terminology and decision records; interview the code domain model, then write each resolved term and decision into CONTEXT and ADR files. Not for remote, credential, publish, deploy, or irreversible changes.

Use when a repository or subsystem needs dead, duplicate, superseded, or generated residue removed across more than one artifact class — code, docs, tests, dependencies, config, workflows, generated files — while preserving observable behavior. Evidence-backed cross-surface purge, not a single enumerated set. Not for deleting an already enumerated record set — use clean-clean-cut.

Use when asked to prune an existing design or codebase until only primitives remain, producing a reusable first-principles map for rewrite or study. Don't use for remote, credential, publish, deploy, or irreversible changes.

Use when a human invokes this skill to build a webhook UI that wakes a bot with server-side secret isolation and an end-to-end probe. Not for unpreviewed or unapproved credential, host, Tailscale, deployment, remote, paid, or irreversible changes.

Use when a user requests an independent Codex or Gemini review of uncommitted code, a branch diff, or a specific commit. Returns findings with exact locations and confidence. Don't use for unconfirmed or unattended external review sends, or for local source mutation.

Use when asked to run /model-retuning [target model or symptom] [corpus path] [bar:<n> consecutive clean runs] to retune a skill corpus for a new model measurement-first. Not for static audits — use deslop for those.

Use when the user mentions mewt, muton, or mutation testing and asks to initialize, scope, estimate, configure, validate, or optimize a campaign before execution. Writes only the TOML config and validates readiness without generating mutants or touching source. Not for running the campaign: use mewt run separately.

Use when a mutation campaign has produced surviving mutants that need triage. Classifies each mutant as false-positive, missing-test, genotoxic, or removable using a coverage map, and proposes which tests to write or dead code to remove. Not for configuring or running campaigns: use mutation-campaign-configuration.

Use when the user asks to detect clever-concealment patterns that obscure real code structure. Scans every file exhaustively and flags each instance with location and severity. Not for refactoring the concealment away: use unleak-abstraction.

Use when asked to optimize code, speed up a path, reduce allocations, repair a regression, or profile a target. Full mode lands one behavior-gated hot-path change proving at least 1.05x speedup; quick mode runs a measure-identify-fix-verify loop with a noise-aware keep/revert and a CI guard. Not for remote, credential, publish, deploy, or irreversible changes.

Use when a product surface must be tested against extreme or hostile worlds rather than collect green checks. Completeness mode classifies every world against stated criteria and reports gaps; break mode names an expected break signal per world, captures proof artifacts, and escalates worlds that survive. Not for interpreting a design dispute — use possible-worlds; not for remote, credential, publish, deploy, or irreversible changes.

Use when a user invokes this skill to generate three targeted questions that prove the author understands how the submitted change affects the existing codebase. Not for reviewing the change — use for pre-review author self-check only.

Use when property-based testing, theorem proving, or formal proof tactics require zero unproven properties. Produces passing proofs and tests, sufficient coverage, and regression tests for every counterexample. Not for remote, credential, publish, deploy, or irreversible changes.

Use when adding or improving property tests for invariants, oracles, parsers, algorithms, data structures, or smart-contract state machines. Encodes the strongest grounded property with domain-aware generators and pinned edge cases. Not for review — use property-test-review.

Use when a generated or shrunk counterexample from a property-based test fails. Classifies whether the defect is in the implementation, the property statement, or the specification, and states the minimal repair action. Not for writing or reviewing tests.

Use when reviewing existing property tests for coverage and defects. Reports tautological, vacuous, assertion-free, reimplemented, weak, or over-filtered tests with evidence, severity, and strongest replacement property. Not for generating tests — use property-test-authoring.

Use when deciding whether PBT fits a code path, which property and library to choose, or whether a design exposes a meaningful property. Returns a verdict with strongest non-vacuous property, domain, strategy seam, and library, or a decline. Not for writing or reviewing.

Use when codebase terminology or a durable technical decision changes. Records resolved terms and qualifying decisions to the project glossary and architecture decision log. Not for shaping domain language — use domain-modeling. Not for remote or irreversible changes.

Use when modernizing APIs, removing compat shims, killing feature flags, or rewriting a subsystem cleanly. Deletes old paths and rewrites every caller to the new contract with zero legacy residue. Not for additive refactors that must preserve the old path.

Use when an implementation has accumulated more workarounds than structure and another patch will not pay. Archives the failed implementation as evidence, carries forward only evidence-earned lessons, and delivers a fresh v0 vertical loop through a predeclared gate. Not for re-deriving a subsystem contract in place: use breaking-driven; not for rewriting one sedimented artifact: use rewrite-clean-v0.

Use when an exact symbol, path, entrypoint, or line range can bound a focused code question or patch proposal under a fixed source budget. Builds a deterministic slice packet and validates one constrained delegation. Not for source changes or broad repository exploration.

Use when a smart-contract codebase needs an evidence-based maturity scorecard and prioritized improvement roadmap. Rates nine categories from code and explicit off-chain evidence, computes an aggregate score, and produces a CRITICAL-to-MEDIUM roadmap. Not for vulnerability auditing, source fixes, or claims about inaccessible processes.

Use when a smart-contract project needs architecture, implementation, dependency, or testing guidance. Produces an evidence-backed assessment with recommendations. Not for maturity — use smart-contract-code-maturity-assessor; not for audit prep — use smart-contract-audit-prep.

Use when implementation must be checked requirement-by-requirement against an authoritative specification, with evidence for each verdict. Handles both standalone spec audits and PR-review spec-drift checks against checked-in spec context. Not for writing or updating specs: use spec-driven-implementation. Not for remote or irreversible changes.

Use when a user invokes a strict-mode validation or verifiable-goals loop setup. Bootstraps strict-mode tooling and per-task GOALS.md scaffolding so an agentic loop can self-verify. Don't use for remote, credential, publish, deploy, or irreversible changes.

Use when about to delete a test on the claim that a new harness, suite, or scenario already covers it. Requires an independent reviewer to validate the per-assertion coverage table and a tree-wide reference scrub before the deletion commit. Not for tests that catch no real bug: use tests-purge-unneeded; not for untracked data or changes without VCS rollback.

Use when a legacy, slow, or duplicate test suite is being cleaned up, a post-refactor sweep is due, or the type system already covers an asserted contract. Not for tests superseded by a new harness — use test-migration-coverage-gate; not for untracked data or changes without VCS rollback.

Use when the user says greenfield this or rescue this codebase, or names a field: dark, red, blue, or brown. Also handles per-subsystem diagnosis. Not for writing specs — use to-spec; not for remote or irreversible changes.

Use when a user names an abstraction leak and wants it sealed as a module seam, configuration option, or explicit override, or deliberately exposed as a named boundary. Measures hidden and wrapper complexity as branching decisions and admits the change only if the wrapper adds less than half the hidden complexity. Not for detecting concealment patterns: use no-hide.

Use when building protocols, workflows, concurrent systems, or lifecycle-heavy state that needs explicit states, transitions, and temporal properties. Defines the state machine, encodes invariants in types, and for high-risk designs runs a TLA+ or Alloy model checker. Not for encoding domain models in types — use type-driven; not for design-by-contract — use contract-driven.

Use when asked to create a project-local executable verification skill, or to repair one whose commands, paths, or assertions have drifted from the project. Writes only inside the verification-skill directory and proves the harness against the live repository. Not for remote or deployed verification, use the project s remote-proof workflow.