Find Failures — AI skill for Claude Code
Find how your AI product fails: review real traces in Eval Studio and group your notes into failure modes.
How to install Find Failures
Installs to ~/.claude/skills/ryanalberts-pmstack-find-failures/SKILL.md
mkdir -p ~/.claude/skills/ryanalberts-pmstack-find-failures && curl -fsSL https://raw.githubusercontent.com/RyanAlberts/pmstack/HEAD/commands/find-failures.md -o ~/.claude/skills/ryanalberts-pmstack-find-failures/SKILL.md Restart Claude Code, or start a new session, for it to be picked up.
What Find Failures does
description: "Find how your AI product fails: review real traces in Eval Studio and group your notes into failure modes" argument-hint: "[folder or trace file]"
Use the Skill tool to run the `pmstack-find-failures` skill from the pmstack plugin with these arguments: $ARGUMENTS. If the skill is not available, tell the user to run /plugin install pmstack@pmstack.
Alternatives in AI
- Deepreasoning — A high-performance LLM inference API and Chat UI that integrates DeepSeek R1's CoT reasoning traces with Anthr 5.4k ★
- Botmux — Bridge Feishu/Lark to AI coding CLIs — Claude Code, Codex, Gemini, OpenCode… every DM, group or topic spawns i 1.2k ★
- Debug Skill — Give your AI agent a real debugger — breakpoints, stepping, variable inspection, and stack traces via CLI 218 ★
Full documentation available on GitHub
View Source RepositoryRelated Skills
Claude Failures
Documented failure modes from working with Claude Code. Post-mortems of real incidents where the AI agent assu
Design AI Feature
Design an AI-powered feature end-to-end — model selection, prompt architecture, eval framework, failure modes,
Codex Plan Review
Adversarial Codex review of a PLAN or design (not code) via the local Codex CLI. Read-only second-model challe
Log Failure
Record an AI mistake in the fixed format — in the current ticket's plan and, if it generalises, in the repo-wi
Silent Failure Scanner
Scan AI outputs for silent failures: placeholders, empty stubs, hedging, no-progress repeats, hallucination ma
Autonomous Debugger
Evidence-driven debugging skill for Claude Code and AI coding agents. Investigates codebases, reproduces failu
Related Agents
Agentic Reviewer
Independent code reviewer for LLM and agent code. Use after implementing a feature, or when asked to review (\
Sme Eval Triage
Triage a golden-set failure from the compliance-SME seat before anyone edits ground truth. Use whenever make e
Decision Pre Mortem
Pre-mortem analysis of a planned decision or project -- imagines failure, traces failure modes to root causes,
