Menu
Build and maintain the benchmark task set for the harness evolution plugin.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends AI agent capabilities with specialized knowledge, workflows, or tool integrations.
Use when creating, scaffolding, or updating the repository AI harness in VS Code. Trigger on requests to bootstrap or refresh instructions, prompts, agents, hooks, skills, or memory layers while preserving useful existing structure and avoiding monolithic always-on context.
Canonical SEA-Forge release workflow for release preparation, validation, branch promotion, tagging, rollback, and release-gate evidence. Use when preparing a release, promoting dev to stage or stage to main, cutting a version tag, validating release readiness, handling hotfix forward-merges, planning rollback, or executing a full end-to-end release. Prefer existing just recipes and evidence-producing gates over ad hoc deployment steps.
Use when SEA work touches specs, generators, manifests, semantic fixtures, regeneration, `src/gen`, or behavior projected from ADR/PRD/SDS/SEA authority.
Use when a SEA context has valid specs and generated contracts but lacks handwritten logic, runtime wiring, persistence, integration, tests, or executable proof.
Create concise, copy-paste-ready bridge prompts between a repo-aware coding agent and an external strategic reasoning agent such as S1 ArchDevAgent. Use when the user wants to offload architecture, research, large refactor planning, migration design, plan critique, implementation-spec drafting, test-strategy design, whole-repo analysis, or diff review to an external agent while keeping repo inspection and implementation in the coding agent. Also use when the user says "use S1", "ask S1", "make a prompt for S1", "bridge this to my external agent", "prepare a context packet", or "validate this S1 response".
| name | harness-benchmark-builder |
| description | Build and maintain the benchmark task set for the harness evolution plugin. |
Curate the benchmark suite used to gate candidate promotion. This skill keeps task coverage balanced across audit, governance, workflow, and held-out regression checks.
scripts/harness_core/benchmarks/tasks/.python ${CLAUDE_PLUGIN_ROOT}/scripts/harness_core/benchmark_runner.py \
--harness-dir $CLAUDE_PROJECT_DIR \
--tasks-dir ${CLAUDE_PLUGIN_ROOT}/scripts/harness_core/benchmarks/tasks \
--run-type baseline \
--dry-run