Stats
Actions
Tags
From superomni
Run a full **harness engineering audit** on the superomni skill framework — evaluate context efficiency, tool action space, evaluation gate coverage, feedback loops, and documentation freshness.
How this command is triggered — by the user, by Claude, or both
Slash command
/superomni:harness-auditThe summary Claude sees in its command listing — used to decide when to auto-load this command
# /harness-audit Run a full **harness engineering audit** on the superomni skill framework — evaluate context efficiency, tool action space, evaluation gate coverage, feedback loops, and documentation freshness. ## Usage ## What Happens Invokes the `harness-engineering` skill, which: 1. **Inventories** the harness (skill count, agent count, preamble size, validation status) 2. **Audits context window pressure** — preamble efficiency + skill bloat 3. **Audits tool action space** — minimalism and composability 4. **Audits evaluation gates** — coverage at every workflow transition 5. **...
Run a full harness engineering audit on the superomni skill framework — evaluate context efficiency, tool action space, evaluation gate coverage, feedback loops, and documentation freshness.
/harness-audit — run full harness audit (all 6 phases)
/harness-audit context — audit context window and preamble efficiency only
/harness-audit gates — audit evaluation gate coverage only
/harness-audit feedback — audit feedback loops only
/harness-audit gc — run garbage collection pass (stale docs, dead commands)
Invokes the harness-engineering skill, which:
A structured Harness Audit Report saved to docs/superomni/harness-audits/harness-audit-[branch]-[timestamp].md.
"Engineers design the system. Agents execute." — OpenAI Harness Engineering
The harness is the product. Running this audit is how you maintain it.
See docs/HARNESS.md for the full harness engineering guide.
npx claudepluginhub wilder1222/superomni --plugin superomni/harness-auditRuns a deterministic repository audit across configurable scopes (repo, hooks, skills, commands, agents) and returns a prioritized scorecard with category scores, failed checks, and top actions.
/harness-auditAnalyzes review agent effectiveness, model routing, and orchestration complexity from metrics data. Produces report on components for simplification or removal, with optional --output path.
/harness-auditAudits the repository against the session-orchestrator harness rubric and reports the overall score, band (healthy/warn/critical), and any categories scoring below 7.
/harness-auditAudits the harness: verifies HARNESS.md against project state by discovering tools/CI/hooks/constraints, detects drift, updates status/badge, reports results and recommendations.
/harness-auditRuns deterministic harness audit on repo (or hooks/skills/commands/agents scopes), producing scorecard with 7 category scores, failed checks, top actions, and suggested skills. Supports text/json formats.