Skip to content

vox-effort-audit API Reference (generated)

Crate: vox-effort-audit · Version: 0.6.0 · Target: Library (lib.rs)

AI-judged audit of git commit history, ranking commits by estimated agent token spend

AI-judged audit of git commit history.

See docs/superpowers/specs/2026-05-28-effort-audit-core-design.md.

DependencyKind
vox-actor-runtimeNormal
vox-configNormal
vox-gitNormal
vox-telemetryNormal
ModuleDescription
config—
hybrid—
judge—
output—
pipeline—
pricing—
range—
shape—
walk—
Dependency Topology, diagram 18 nodes, 7 edges: vox-effort-audit → vox-actor-runtime; vox-effort-audit → vox-config; vox-effort-audit → vox-git; vox-effort-audit → vox-telemetry; vox-cli → vox-effort-audit; vox-cli-research → vox-effort-audit; vox-effort-route → vox-effort-auditvox-effort-auditvox-actor-runtimevox-configvox-gitvox-telemetryvox-clivox-cli-researchvox-effort-route
Diagram source
graph {
direction: right
node vox_effort_audit: "vox-effort-audit"
node vox_actor_runtime: "vox-actor-runtime"
vox_effort_audit -> vox_actor_runtime
node vox_config: "vox-config"
vox_effort_audit -> vox_config
node vox_git: "vox-git"
vox_effort_audit -> vox_git
node vox_telemetry: "vox-telemetry"
vox_effort_audit -> vox_telemetry
node vox_cli: "vox-cli"
vox_cli -> vox_effort_audit
node vox_cli_research: "vox-cli-research"
vox_cli_research -> vox_effort_audit
node vox_effort_route: "vox-effort-route"
vox_effort_route -> vox_effort_audit
}
Consumer CrateRole
vox-cliDirect workspace dependent
vox-cli-researchDirect workspace dependent
vox-effort-routeDirect workspace dependent
KindSymbolSignatureSourceDescription
structconfig::EffortAuditConfigpub struct EffortAuditConfigcrates/vox-effort-audit/src/config.rs#L8-L32—
structconfig::JudgeConfigpub struct JudgeConfigcrates/vox-effort-audit/src/config.rs#L35-L43—
enumhybrid::MeasuredCostpub enum MeasuredCostcrates/vox-effort-audit/src/hybrid/mod.rs#L9-L22—
structjudge::JudgeOutcomepub struct JudgeOutcomecrates/vox-effort-audit/src/judge/mod.rs#L17-L24—
enumjudge::JudgeStatuspub enum JudgeStatuscrates/vox-effort-audit/src/judge/mod.rs#L27-L31—
enumjudge::SkipReasonpub enum SkipReasoncrates/vox-effort-audit/src/judge/mod.rs#L34-L37—
traitjudge::Judgepub trait Judge: Send + Synccrates/vox-effort-audit/src/judge/mod.rs#L40-L43—
structjudge::MockJudgepub struct MockJudgecrates/vox-effort-audit/src/judge/mod.rs#L46-L49Deterministic in-memory judge for tests.
structjudge::LlmJudgepub struct LlmJudgecrates/vox-effort-audit/src/judge/mod.rs#L96-L100Real judge wired through vox_actor_runtime::llm with the chosen model. All LLM I/O goes through vox_actor_runtime::llm::infer_with_retry. The model id is resolved upstream (orchestrator model registry) and passed in as resolved_model. The judge surface is intentionally vendor-agnostic — no provider hostnames or SDKs leak in here. See AGENTS.md §Model-Agnostic LLM Boundary.
structoutput::FindingRowpub struct FindingRowcrates/vox-effort-audit/src/output/mod.rs#L11-L23The top-level shape of one line in findings.jsonl. Stable schema_version=“1.0”.
structoutput::JudgeMetapub struct JudgeMetacrates/vox-effort-audit/src/output/mod.rs#L26-L32—
structpipeline::RunSummarypub struct RunSummarycrates/vox-effort-audit/src/pipeline.rs#L39-L44Summary returned to the caller after run finishes. Used by the CLI (F1) to print a one-liner without re-reading manifest.json.
fnpipeline::runpub async fn run(crates/vox-effort-audit/src/pipeline.rs#L68Run one vox audit effort pass. Composition: 1. Resolve the commit range (range::resolve). 2. Walk commits, computing diffs + shape features. 3. For each commit: resolve hybrid measured cost, check the per-run token budget, call the judge, emit a FindingRow line to findings.jsonl. 4. Render report.md + manifest.json. transcript_dir_override lets tests point at a fixture transcript tree; None means use cfg.transcript_dir. CLI bookend overrides (F1): - since_override: when Some, overrides cfg.default_since (the --since flag). Accepts the same forms as the TOML key ("30 days ago", "7d", git refs, dates). - until_override: when Some, overrides the default HEAD upper bound (the --until flag). Same form rules as since_override. cfg.limit, when set, caps the number of judged commits — commits past the cap are emitted with JudgeStatus::Skipped(LimitReached) so the manifest.json commits_in_range count still reflects the full walk.
fnpipeline::run_with_overridespub async fn run_with_overrides(crates/vox-effort-audit/src/pipeline.rs#L92Same as [run] but accepts explicit --since / --until overrides from the CLI. Internal entry-point — keeps the public [run] signature stable for callers that don’t need to wire bookend overrides (tests, library use).
structpricing::ModelRatespub struct ModelRatescrates/vox-effort-audit/src/pricing.rs#L8-L13USD-per-1K-token rates for the resolved judge model.
fnpricing::cost_usdpub fn cost_usd(&self, prompt_tokens: u64, completion_tokens: u64) -> Option<f64>crates/vox-effort-audit/src/pricing.rs#L17-L25Real cost in USD, or None when pricing is unknown (NOT a fake 0.0).
enumrange::CommitRangepub enum CommitRangecrates/vox-effort-audit/src/range.rs#L7-L12—
enumrange::RangeErrorpub enum RangeErrorcrates/vox-effort-audit/src/range.rs#L15-L18—
fnrange::parse_durationpub fn parse_duration(s: &str) -> Result<Duration, RangeError>crates/vox-effort-audit/src/range.rs#L25-L70Parses a duration string of the form NNd, NNh, or NNw (N = ASCII digits), or the long forms "N days ago" / "N hours ago" / "N weeks ago". The compact digits+suffix form requires the prefix to be ALL ASCII digits — refs like feature-2d or abc123d MUST NOT parse as durations.
fnrange::resolvepub fn resolve(crates/vox-effort-audit/src/range.rs#L76Resolves CLI args + config default into a CommitRange. since_arg and until_arg are the raw --since/--until strings. If neither parses as a duration, both are treated as git refs.
fnrange::duration_cutoffpub fn duration_cutoff(now: DateTime<Utc>, d: Duration) -> DateTime<Utc>crates/vox-effort-audit/src/range.rs#L94-L96For SinceDuration, the wall-clock cutoff at run time.
structshape::ShapeFeaturespub struct ShapeFeaturescrates/vox-effort-audit/src/shape.rs#L8-L18—
enumshape::CommitKindpub enum CommitKindcrates/vox-effort-audit/src/shape.rs#L22-L32—
fnshape::featurespub fn features(rec: &CommitRecord) -> ShapeFeaturescrates/vox-effort-audit/src/shape.rs#L34-L79—
structwalk::CommitRecordpub struct CommitRecordcrates/vox-effort-audit/src/walk.rs#L19-L30—
structwalk::FileChangepub struct FileChangecrates/vox-effort-audit/src/walk.rs#L33-L37—
enumwalk::WalkErrorpub enum WalkErrorcrates/vox-effort-audit/src/walk.rs#L40-L45—
fnwalk::iter_commitspub fn iter_commits(crates/vox-effort-audit/src/walk.rs#L48Iterates commits in the range, newest-first.