Use this for one candidate; use check_executables for several, and do not use it for fuzzy
    discovery—use search_similar_executables. name is case-sensitive and is not normalized;
    scope filters providers with AND semantics before the collision decision, so a scoped miss is
    non-exhaustive. The tool makes no live registry call and never runs or installs software; a
    hit returns providers, while a miss is clear only when coverage_scope is exhaustive.
    

    Use this for multiple candidates; use check_executable for one, and do not use it to rank
    similar spellings—use search_similar_executables. One scope filters every name with AND
    semantics; each name is evaluated independently and results, including duplicates, preserve
    input order. The tool makes no live registry call and never runs software; treat misses as
    unknown unless coverage_scope is exhaustive, which a supplied scope never is.
    

    Use this after lookup for one known target; use resolve_installations for several commands,
    and do not use it for name availability—use check_executable. environment maps its package
    manager to a provider ecosystem before preferences.package narrows candidates; omitted target
    facts are returned in required_context rather than guessed. The tool never executes argv or
    probes package availability; status and verification distinguish ready, ambiguous,
    unsupported, unverified, missing-index, and incomplete-context outcomes.
    

    Use this when all commands share an environment and preferences; use resolve_installation
    for one, and do not use it for availability—use check_executables. The shared environment
    selects an ecosystem before preferences.package narrows each command independently; results,
    including duplicates, preserve input order. The tool never executes argv or probes package
    availability, and each result reports its own status, required_context, and verification.
    

    Use this to inspect package, ecosystem, source, version, and provenance; use check_executable
    for collision decisions, and do not treat a null record as availability. name is not
    normalized; scope applies AND filters to providers before deciding whether record is null, so
    a known command can disappear when no provider matches. The tool reads one fixed snapshot and
    performs no live lookup; interpret null with coverage_scope.
    

    Use this for exploration; use check_executable for exact collisions and
    search_similar_executables for confusable names. All active filters intersect; ecosystem and
    scope.ecosystem must both match, so conflicting values return none, and limit applies only
    after deterministic case-insensitive sorting with case-sensitive tie-breaking. Do not use an
    empty result to infer availability: the tool makes no live registry call and fails instead of
    returning partial data when a required derived index is missing or corrupt.
    

    Use this after an exact check for typo risk; use check_executable for exact existence, and do
    not substitute this fuzzy result. name seeds case-folded trigram candidates, scope filters
    them before scoring, and limit applies after ranking by edit distance, trigram similarity,
    then name. The tool makes no live registry call; an empty match list covers only the fixed
    snapshot and optional scope, not global absence.
    

    Call this with no parameters before interpreting a miss; use check_executable to look up a
    command, and do not use coverage metadata as an existence result. negative_lookup becomes
    exhaustive only when every declared source succeeded with exhaustive, unrestricted coverage;
    otherwise absence remains unknown. The tool reads fixed metadata without probing registries,
    and freshness may be unavailable because its partial scan is stored separately.
    

    Use this after an existence check; use assess_executables for several names, and do not use it
    instead of check_executable. scope filters providers before signals are aggregated; release
    age is measured against request time with a 365-day threshold, while usage counts can still
    establish activity. The tool makes no live registry call; unknown or insufficient_evidence
    means missing snapshot evidence, not that the name is safe.
    

    Use this to compare candidates; use assess_executable for one, and do not use it instead of
    check_executables. One scope filters every name before per-name signal aggregation; release
    age uses request time and results, including duplicates, preserve input order. The tool makes
    no live registry call; unknown or insufficient_evidence means missing snapshot evidence, not
    safety, so retain each result's existence and coverage context.