pride

Review git diffs and uncommitted changes for defects and misleading claims.

3|1|Updated Apr 4, 2026
One-click install
npx skills add https://github.com/epologee/leclause-skills --skill pride
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: pride
Source: https://github.com/epologee/leclause-skills/tree/main/packages/autonomous/skills/pride
Command: npx skills add https://github.com/epologee/leclause-skills --skill pride

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Pride enforces a skeptical, human-like review of artefacts produced by autonomous agents so that subtle defects, misleading claims, and confidence laundering do not slip past CI or automated checks. It prevents an automatic "green is good" mindset by forcing explicit, evidence-backed findings and concrete fixes before work is considered done.

Core Features & Use Cases

  • Contrarian code review: Reviews git diffs and uncommitted changes to find duplicate fixes, type smells, race conditions, and unsafe shell usage.
  • Prose and media review: Evaluates research briefs, plans, PR descriptions, and media captions for confidence laundering, unsourced claims, and over-claims.
  • Gated autonomy: Runs as an INSPECT-phase gate inside a rover loop or is invocable manually to produce actionable findings, set the phase back to DRIVE, or print results to a user conversation.
  • Practical workflows: Use it to check a branch before pushing, audit generated documentation for unsupported assertions, or review mixed artefacts that combine code and narrative.

Quick Start

Run the pride review on the current branch including uncommitted changes so the subagent inspects the diff and returns blunt findings with concrete fixes.

Frequently Asked Questions about pride

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I perform a contrarian code review on a git diff to find overlooked defects?

A contrarian code review examines git diffs to surface overlooked defects, duplicate fixes, and unsafe shell usage. It enforces skeptical, evidence-backed findings with concrete remediation steps instead of assuming passing checks mean the code is safe.

What is confidence laundering in generated documentation and how do I check for it?

Confidence laundering is when generated prose masks unsupported assertions or over-claims as factual. You check for it by applying a skeptical review to research briefs and documentation, demanding explicit evidence and flagging misleading claims before publication.

Can I review uncommitted changes and generated media captions in the same autonomous inspection phase?

Yes, you can review uncommitted changes and generated media captions together. The inspection process collects diffs and artefact text simultaneously, evaluating mixed code and narrative outputs to find both type smells and unsourced claims.

How do I review a git branch before pushing to catch race conditions and type smells?

Reviewing a branch before pushing involves running a skeptical inspection on the current git range. This surfaces race conditions, type smells, and unsafe shell usage, returning per-file findings with concrete fixes to apply before pushing.

Does automated quality assurance catch misleading claims in generated research briefs?

Automated quality assurance often misses misleading claims in generated research briefs because it checks for passing tests rather than factual accuracy. A contrarian review process explicitly audits prose for unsourced claims and confidence laundering that automated checks overlook.

What are the limitations of relying on automated checks instead of a contrarian review for code diffs?

Automated checks often adopt a green-is-good mindset, missing subtle defects like race conditions and confidence laundering. A contrarian review is limited by token-aware truncation for large payloads, but it enforces the evidence-backed findings that automated checks lack.