harness

Sprint feature assignment planner. Analyze the dependency graph and team workload from docs/features.json, output the optimal owner assignment plan + a ready-to-run sprint-kickoff.sh script.

Perform a health audit on the current project's Harness system, generating a score report and optimization recommendations

Generate a canary deployment runbook with risk assessment, staged rollout plan, rollback triggers, and observability checklists. Outputs to deploy/ directory.

Save key decisions and progress from the current session to documentation, for cross-session handoff of long tasks

Initialize the full AI Agent Harness engineering system for the current project

Code quality Inferential Sensor. Invoke in the following scenarios: code quality review before PR merge, architecture convention compliance checks, design evaluation of new business logic, baseline assessment before refactoring. Division of responsibilities with security-reviewer: this Agent focuses on quality/architecture, security-reviewer focuses on security vulnerabilities.

Long-cycle multi-session coding tasks. Invoke when implementing a set of features across multiple sessions: new feature iteration, large-scale refactoring, multi-module collaborative development. Characteristics: task expected to span more than one session, has a features.json feature checklist, requires strict "one feature at a time" constraint to prevent context anxiety. Differs from the main Agent: enforces startup checklist, cross-session state handoff, and a mandatory two-phase Review after each feature (spec compliance → code quality).

Lightweight codebase exploration. Invoke when investigating a specific module, function, or pattern to avoid polluting the main thread context with large file reads. Applicable scenarios: understanding auth/authorization flows, finding reusable utility functions, investigating all implementations of an interface, cross-module dependency analysis.

Professional security code review. Invoke in the following situations: pre-commit review, new authentication/authorization logic, external API integrations, user input handling.

> **Source**: OpenSpec `/opsx:archive` + Handbook S K.6 "Documentation Sync Agent" + Handbook S 2.3 "Structured Handoff Artifacts"

Health check and optimization for existing project Harness setups. Activate when the user mentions "check Harness", "optimize CLAUDE.md", "Agent keeps making mistakes", "Harness health", "audit Harness", "evaluate AI coding environment", "harness audit", "check Agent config", "why won't the Agent follow instructions", "improve Agent effectiveness", "add Harness to existing project", or "legacy optimization". Also use this Skill when the user complains that Agent behavior deviates from expectations, the Agent repeatedly makes the same mistakes, or the project has been around for a while but lacks a systematic Harness framework — diagnose and improve.

You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.

> Generate a structured deployment runbook with risk-based canary stages, rollback triggers, and observability checklists.

Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies

6 hooks across 4 events