/verify

> Test L0 abstract fidelity -- do compressed summaries match actual content?

170 stars

byMiosa-osa

View on GitHub Installation ↓

Best use case

/verify is best used when you need a repeatable AI agent workflow instead of a one-off prompt.

> Test L0 abstract fidelity -- do compressed summaries match actual content?

Teams using /verify should expect a more consistent output, faster repeated execution, less prompt rewriting.

When to use this skill

You want a reusable workflow that can be run more than once with consistent structure.

When not to use this skill

You only need a quick one-off answer and do not need a reusable workflow.
You cannot install or maintain the underlying files, dependencies, or repository context.

Installation

Claude Code / Cursor / Codex

$curl -o ~/.claude/skills/verify/SKILL.md --create-dirs "https://raw.githubusercontent.com/Miosa-osa/canopy/main/library/skills/knowledge/verify/SKILL.md"

Manual Installation

Download SKILL.md from GitHub
Place it in .claude/skills/verify/SKILL.md inside your project
Restart your AI agent — it will auto-discover the skill

How /verify Compares

Feature / Agent	/verify	Standard Approach
Platform Support	Not specified	Limited / Varies
Context Awareness	High	Baseline
Installation Complexity	Unknown	N/A

Frequently Asked Questions

What does this skill do?

> Test L0 abstract fidelity -- do compressed summaries match actual content?

Where can I find the source code?

You can find the source code on GitHub using the link provided at the top of the page.

SKILL.md Source

# /verify

> Test L0 abstract fidelity -- do compressed summaries match actual content?

## Usage
```
/verify [--sample <n>]
```

## What It Does
Samples context files, reads their L0 abstract (compressed ~100 token summary) and their full content, then evaluates whether the abstract accurately represents the content. Reports fidelity scores and flags abstracts that have drifted from their source.

## Implementation
Runs: `cd engine && mix optimal.verify [--sample <n>]`

Process:
1. Sample N context files (default: 10).
2. Load L0 abstract for each.
3. Load full content for each.
4. Compare: does the abstract capture the key facts?
5. Score fidelity (0-1) per file.
6. Flag any below threshold (< 0.7).
7. Suggest re-generation for drifted abstracts.

## Examples
```bash
# Test 10 random contexts
/verify

# Test a larger sample
/verify --sample 20
```

## Command