/elevate
Reference for the /elevate command: invocation, arguments, inputs and outputs.
Aggressive, zero-trust, open-loop MASTER repository-elevation command. DETECTS across every auditable dimension by dispatching the report-only /audit fortress (reusing its eleven passes, never re-deriving them), then runs independent critique agents over the elevation dimensions no audit covers (storytelling, naming uniformity, cross-sync, consolidation/modularity, aesthetics, visualization/tabulation, extensibility), folder by folder, file by file, line by line. Lifts each surface to current SOTA per rules/sota-elevation.md, remediates with the most fitting action (destruction included; every destructive step confirmation-gated), propagates each change across the reference graph per rules/propagation.md, and ends trace-free and production-ready by handing off to /freshify. Orchestration runs through the workflow harness (skills/workflow/SKILL.md); no model tier or effort is preset. WARNING: a full run is a relentless multi-pass undertaking — bound it with --dimensions / --max-passes for a scoped run.
Invocation
/elevate [path/to/repo/] [--dimensions all|<list>] [--max-passes N] [--open-loop] [--autonomous] [--verify-panel N] [--strict]The definition sets disable-model-invocation: true, so a harness that honors that key starts this command only when you type it.
Pipeline position
/elevate is the aggressive elevation pass that precedes the freshening close, and the remediation consumer the report-only /audit routes its HIGH findings to. It consumes the target repository's full working surface, emits an elevated cross-coherent tree plus an elevation report, then hands off to /freshify for the trace-free culmination. It is distinct from /audit (report-only multi-dimension detection), /fortress (the security/production-scoped detect → remediate → re-audit → gate loop, which routes its broad lifts here), and /release-readiness (a single READY / BLOCKED verdict): /elevate is the broad, open-loop, whole-repo critique-and-remediate to SOTA. The elevation here is the substrate the freshening core finalizes.
Inputs
The target repository's entire working surface — every file, folder, and package; every root file; the README files, AGENTS.md, CLAUDE.md, the website / portal and its source, the documentation tree, the test suite, the examples, the assets and logos, the .github/ folder, the source-forge metadata, the CHANGELOG, and every narrative and style surface — plus /audit's verified findings (the technically-auditable detection front), the host's ratified conventions discovered per rules/host-discovery.md, and current SOTA authoritative sources accessed per the source-access contract below.
| Argument | Type | Required | Description |
|---|---|---|---|
path/to/repo/ | Path | Yes | Root directory of the target repository. The command refuses execution when no elevation surface resolves. |
--dimensions all|<list> | Enum/List | No | Scope the cross-dimension critique to the named axes (comma-separated) or, with all, the full axis list. Default all. The audit-detected dimensions accept the same names the eleven first-class passes use (e.g. security-audit,perf-audit). |
--max-passes N | Integer | No | Cap the open-loop pass count at N per the iteration-safety contract. Without the flag, the loop runs to convergence under the bounded retreat protocol. |
--open-loop | Flag | No | Enable dynamic open-loop "dreaming" orchestration — independent critique agents re-fan across nuance classes between passes per rules/agent-orchestration.md rather than running a single fixed wave. |
--autonomous | Flag | No | Opt into continuous pass-to-pass advancement without a per-pass-boundary halt. Default: halt at each pass boundary and confirm, per rules/agnostic-posture.md + rules/context-management.md §4A. Destructive remediations stay per-target gated even under this flag. |
--verify-panel N | Integer | No | Refute-by-default critics per surviving finding in the bespoke-critique fan-out (default: 3). |
--strict | Flag | No | Promote every advisory elevation finding to a blocking finding. Under --strict, elevation is complete only when zero sub-SOTA surfaces and zero open critiques remain. |
Outputs
The elevated working tree (in-place by default), cross-coherent at SOTA across every dimension, and an elevation report enumerating every nuance inspected, every critique verdict (with the /audit findings it consumed cited, not re-derived), every remediation with its action and rationale, every surfaced elevation opportunity, the per-dimension attestation, the per-axis attestation against the seven-axs-of-breadth taxonomy, the disclosure ledger of every amendment / extension / refinement per rules/disclosure-ledger.md, and the SOTA-elevation disclosure markers per rules/sota-elevation.md.
Next step
Invoke /freshify against the elevated tree to culminate the project trace-free and production-ready — /elevate lifts every surface to SOTA, and the freshening core finalizes the clean, current-version-only close per rules/production-ready-prs.md.
Source
Generated from src/apothem/commands/elevate.md, the command definition every harness installs.