skills.rest
SamyakJhaveri
SamyakJhaveriCOMMUNITY
Explore all skills by this author →
GitHub

session-critique

Critique session outputs with adversarial, decision-aware rigor before committing.

Updated Jul 2, 2026
Search Tags:#workflow#quality-gate#code-review#adversarial-review#session-critique#decision-adherence
One-click install
npx skills add https://github.com/SamyakJhaveri/loam --skill session-critique
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: session-critique
Source: https://github.com/SamyakJhaveri/loam/tree/main/seed/.claude/skills/session-critique
Command: npx skills add https://github.com/SamyakJhaveri/loam --skill session-critique

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Reviews and critiques a session's completed work before committing, acting like a senior software engineer and AI researcher to ensure decisions, repo rules, and maintainability are respected.

Core Features & Use Cases

  • Surface high-impact issues (HIGH/MEDIUM) and propose authorizations for fixes
  • Enforce a phase-driven workflow (ANALYZE -> FIX -> Elegance Gate -> Handback) with serial and batch handling
  • Support an advisor-pattern team (advisor, self-critic, code-reviewer) for separation of concerns and escalations

Quick Start

Invoke /session-critique after finishing a session to generate adversarial findings before committing.

Frequently Asked Questions about session-critique

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

FAQPage Schema
How do I run an adversarial code review before committing multi-file changes?▼

Adversarial code review assesses multi-file session outputs using a four-phase workflow (ANALYZE, FIX, Elegance Gate, Handback) with role-separated subagents to surface HIGH/MEDIUM issues before committing.

What is a decision-aware session critique for code and data changes?▼

Decision-aware session critique evaluates completed work against a formal decision ledger and repo rules, acting as a senior engineer to ensure maintainability and decision adherence across code, data, and reports.

How do I batch low-priority code review nits separately from high-impact issues?▼

Batch low-priority nits by applying severity-based remediation, which surfaces HIGH/MEDIUM issues for immediate authorization while batching LOW nits until you explicitly authorize fixes.

Can I use an advisor pattern to separate concerns during a code review?▼

Yes, the advisor pattern supports separation of concerns using a team structure (advisor, self-critic, code-reviewer) to handle escalations and enforce a phase-driven workflow during the critique.

What's the best way to enforce a quality gate on AI-generated session outputs?▼

Enforce a quality gate by running an Elegance Gate phase after fixes, performing adversarial rigor to ensure outputs respect repo rules and decisions before final handback.

When should I not use a formal decision ledger for session critique?▼

Avoid a formal decision ledger for trivial, single-line changes, as this workflow targets multi-file changes across code, data, and reports requiring strict severity-based remediation.

Related Skills

Explore More Software Engineering Skills→
jason660519jason660519

plan-review

Critique implementation plans with failure-mode mapping and mandatory diagrams.

Community
Advanced
dhamijadhamija

code-quality

Apply a structured checklist to review code for correctness, readability, security, performance, and maintainability.

Community
Basic
troykellytroykelly
10

security-reviewer

Execute the security-reviewer subagent for code vulnerability detection.

Community
Basic
outcomeengoutcomeeng
1

reviewing-python

Review Python codebases adversarially and issue APPROVED or REJECTED verdicts.

Official
Advanced