@alchemiststudios/pi-harness-skills
Harness engineering skills package for pi
Package details
Install @alchemiststudios/pi-harness-skills from npm and Pi will load the resources declared by the package manifest.
$ pi install npm:@alchemiststudios/pi-harness-skills- Package
@alchemiststudios/pi-harness-skills- Version
0.1.1- Published
- Mar 17, 2026
- Downloads
- 55/mo · 14/wk
- Author
- tunahorse21
- License
- MIT
- Types
- skill
- Size
- 94.8 KB
- Dependencies
- 0 dependencies · 0 peers
Pi manifest JSON
{
"skills": [
"./skills"
]
}Security note
Pi packages can execute code and influence agent behavior. Review the source before installing third-party packages.
README
@alchemiststudios/pi-harness-skills
Harness Engineering skill pack for pi: structured Research → Plan → Execute → QA workflows, plus mapping and ast-grep setup skills.
Why this package
- Operational workflow skills: run consistent research/planning/execution/qa loops.
- Reproducible artifacts: phases write to
.artifacts/research,.artifacts/plan,.artifacts/execute. - Practical tooling included: structural scanning, symbol indexing, dependency tracing, ast-grep setup.
Install
npm
pi install npm:@alchemiststudios/pi-harness-skills
Git
pi install git:github.com/alchemiststudiosDOTai/pi-harness-skills@v0.1.0
Local path
pi install /absolute/path/to/pi-harness-package
Included skills
| Skill | Purpose |
|---|---|
research-phase |
Factual codebase mapping + structural analysis outputs |
plan-phase |
Convert research into execution-ready implementation plans |
execute-phase |
Execute plans with gates, atomic commits, and a single execution log |
qa-from-execute |
QA changed areas only, without modifying source code |
harness-map |
Map a repo’s harness layers and operator surfaces |
ast-grep-setup |
Bootstrap ast-grep rules and rule tests |
agents-md-mapper |
Build/refresh evidence-grounded AGENTS.md |
differential-session-runner |
Durable differential debugging session workflow |
Quick usage
/skill:research-phase
/skill:plan-phase
/skill:execute-phase .artifacts/plan/<plan-file>.md
/skill:qa-from-execute .artifacts/execute/<execute-log>.md
Or let pi auto-load the matching skill from the prompt.
Requirements
- pi coding agent
- Bash environment for included helper scripts (used by
research-phase)
License
MIT