Maps repo harness layers: canonical check commands, local/CI gates, architecture boundaries, structural rules, behavioral verification, docs ratchets, evidence workflows, operator surfaces. Use to understand change safety.
From harness-engineeringnpx claudepluginhub alchemiststudiosdotai/harness-engineeringThis skill is limited to using the following tools:
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Searches prompts.chat for AI prompt templates by keyword or category, retrieves by ID with variable handling, and improves prompts via AI. Use for discovering or enhancing prompts.
Executes implementation plans in current session by dispatching fresh subagents per independent task, with two-stage reviews: spec compliance then code quality.
Map the repository's actual harness: the mechanical checks, policies, workflows, and artifacts that make change safe.
This skill is narrower than generic codebase research. It is specifically for answering questions like:
check actually run?"Map the harness as implemented, not as imagined.
Prefer:
Avoid:
A repo harness usually includes some or all of these layers:
just check, make check, task check, npm test, etc.AGENTS.md, .codex/, environment files, repo-local skills, slash commandsUse this skill when the user asks to:
Inspect common entrypoint files first:
justfileMakefilepackage.jsonpyproject.tomlCapture:
Inspect CI workflows:
.github/workflows/*.yml.gitlab-ci.yml, buildkite, etc.)Capture:
fail-fast is enabledLook for:
Capture:
Look for:
sgconfig.yml, ast-grep rule directoriesCapture:
Look for:
Capture:
Look for:
Capture:
Look for:
Capture:
Look for:
AGENTS.md.codex/environments/*.codex/skills/*Capture:
Write a research artifact to:
memory-bank/research/YYYY-MM-DD_HH-MM-SS_<repo>-harness-map.md
Recommended structure:
---
title: "<repo> – Harness Map"
phase: Research
date: "YYYY-MM-DD HH:MM:SS"
owner: "<agent_or_user>"
tags: [research, harness, <repo>]
---
## Canonical Entry Point
- `path:line-line` → command and subcommands
## Harness Layers
### Layer 1: Local checks
### Layer 2: Architecture boundaries
### Layer 3: Structural rules
### Layer 4: Behavioral verification
### Layer 5: Docs ratchet
### Layer 6: CI matrix
### Layer 7: Evidence workflow
### Layer 8: Operator surface
## Source Index
- `path:line-line` → what this file contributes
## Observed Command Chain
- ordered list of checks from the main command
Your harness map must:
## Canonical Entry Point
- `justfile:22-29` defines `check *args:` and runs Ruff, Import Linter, ty, docs checks, ast-grep, pytest, and Zig checks.
## Layer 2: Architecture Boundaries
- `pyproject.toml:80-110` defines four Import Linter forbidden contracts.
## Layer 6: CI Matrix
- `.github/workflows/ci.yml:13-79` runs seven matrix tasks with `fail-fast: false`.
The repo appears to care about quality and uses several tools.
It has some tests and some linting.
Common next steps after this skill:
plan-phase if the user wants to add or improve a harness layer