An incident log for recording and reviewing cases where an AI assistant gives wrong advice or takes an incorrect action. It groups failures such as invented facts, forgotten context, outdated information, bias, and overconfidence.

Use it to record what happened, the expected result, impact, category, and supporting evidence, then analyse recurring failures after several incidents.

It turns AI mistakes into searchable evidence instead of letting them be forgotten. Reviewing patterns can show where the assistant needs safer limits or better information.

ai-incident — Command for Claude Code by gonzalezpazmonica

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Clone the repo

git clone --depth 1 https://github.com/gonzalezpazmonica/savia

Made for: Claude Code.

Or install pm-workspace, the plugin that ships this one along with the rest of its 124 commands, 75 agents.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/commands/gonzalezpazmonica/savia/ai-incident/github.svg)](https://agentmods.dev/commands/gonzalezpazmonica/savia/ai-incident)

Your own site

<a href="https://agentmods.dev/commands/gonzalezpazmonica/savia/ai-incident"><img src="https://agentmods.dev/badge/commands/gonzalezpazmonica/savia/ai-incident/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/commands/gonzalezpazmonica/savia/ai-incident"><img src="https://agentmods.dev/badge/commands/gonzalezpazmonica/savia/ai-incident.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 17 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 848 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5.1	$0.00017	$0.00848
Opus 5	$0.00009	$0.00424
Sonnet 5	$0.00003	$0.00170
Haiku 4.5	$0.00002	$0.00085

Measured 5d ago against content hash be17f0ceaf35, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

ai-incident scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

Origin

Copies of this mod

1 near-identical copy found in the catalogue:

ai-incident — 100% identical, 0 lines differ

.claude/commands/ai-incident.md · 122 lines

How it starts

The opening of the file, as written. The whole thing — 122 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/ai-incident

🦉 Los errores de Savia son datos. Aprendemos de ellos para mejorar.

Registrar, categorizar y analizar incidentes donde las recomendaciones o acciones de Savia fallaron o fueron incorrectas.

Categorías de Incidentes

BIAS — Savia favoreció un resultado sin justificación objetiva
HALLUCINATION — Savia inventó datos o asumió hechos sin verificar
CONTEXT-LOSS — Savia olvidó o ignoró información crítica
OUTDATED — Savia usó datos desactualizados
BOUNDARY-VIOLATION — Savia excedió sus límites definidos
CONFIDENCE-MISMATCH — Confianza mostrada alta pero resultado incorrecto

Flujo de Registro

/ai-incident new

Savia presenta formulario interactivo:

¿Qué sucedió? — descripción breve del incidente
¿Cuándo? — fecha aproximada
¿Qué recomendó Savia? — recomendación original
¿Qué era lo esperado? — resultado correcto
¿Impacto? — bajo/medio/alto/crítico
¿Categoría? — bias/hallucination/context-loss/outdated/boundary-violation/confidence-mismatch
Detalles adicionales — evidencia o contexto

Análisis Automático

Tras 5+ incidentes registrados:

/ai-incident analyze

Genera:

Estadísticas: total, resueltos, abiertos, tasa
Top categorías: frecuencia de cada tipo
Tendencias: patrones (ej: "sprint-planning es el comando más propenso a errores")
Recomendaciones: acciones para mejorar (cargar datos faltantes, reducir confianza, etc.)

Comandos

/ai-incident list [--proyecto] [--categoría] [--días N]
/ai-incident view {id}
/ai-incident search "{texto}"
/ai-incident analyze [--últimos N]
/ai-incident export [--formato csv|json|md]

Integración con AI Safety

Los incidentes informan automáticamente:

Recalibración de confianza: si recomendaciones de tipo X tienen alto % de incidentes, bajar confianza
Actualización de límites: si un límite se viola frecuentemente, considerarlo
Mejora de context-map: si hay context-loss recurrente, cargar más datos
Alertas: "Has tenido 3 incidentes en asignaciones — Savia pedirá validación extra"

Read the full file on GitHub · 122 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

5d ago First seen · 122 lines · 17 tokens per session scan A be17f0ceaf35

Subscribe to this mod's changes

ai-incident is a command published in the GitHub repository gonzalezpazmonica/savia (50 stars, last pushed yesterday), licensed MIT. It adds 17 tokens to every session and 848 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-06.

Building agents Claude Code Debugging codex harness opencode

root-cause

Use when any test fails, bug appears, or behaviour surprises you, before proposing a fix - find the cause and prove it, by reading real evidence, tracing bad values back to their origin, comparing against a working case, and testing one hypothesis at a time.

hamr0/liteagents · 52 tokens

ai-incident