jwalsh / jwalsh/cprr

feat: add `cprr measure` command for typed evidence with quality levels

Open
#8 0 comments 0 reactions 0 assignees View on GitHub
cprr enhancement
Dominant language
Go
Stars
0
Forks
0
PR merge metrics
No merged PRs in 30d

Description

## Summary

Implement `cprr measure` as a structured evidence command with quality levels.

## Motivation

SKILL.md defines evidence quality levels but the tool has no way to record them:

| Level | Description | Example |
|-------|-------------|---------|
| Anecdotal | Single observation | "Seemed faster" |
| Measured | Quantified observation | "p99: 180ms" |
| Statistical | Significance tested | "p=0.03, n=10000" |
| Replicated | Multiple trials | "3 runs, consistent" |

Currently `cprr evidence "text"` stores flat strings. Agents have attempted to invoke `cprr measure --id --result --status` (found in emacs-mcp-maximalist logs) — a command that doesn't exist but arguably should.

## Proposed Behavior

```bash
cprr measure --level measured --result "p99: 180ms, n=1000"
cprr measure --level statistical --result "p=0.03, n=10000" --status confirmed
```

- `--level` (anecdotal|measured|statistical|replicated) — defaults to anecdotal
- `--result` — the observation text (required)
- `--status` — optionally advance state in one command (checks guards)

This combines evidence recording + optional state transition, which is the most common two-step pattern agents perform.

## Backward Compatibility

`cprr evidence` continues to work as-is (flat text, no level). `cprr measure` is the structured alternative.

Contributor guide

Open the contributing guide

Research direction

Start with SKILL.md and the existing `cprr evidence` command to trace how evidence is stored and how state-transition guards are applied. Add `cprr measure` with the four levels, required `--result`, optional guarded `--status`, and unchanged `cprr evidence` behavior; verify the documented examples and backward compatibility.

Written by the indexing model from the issue text.

Assessment

Tech stack
go
Domain
cli
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
57/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.