infra-auditor

infra-auditor is an agent for Claude Code from Freespirits/claudeguard-il. It costs 185 tokens per session (4,032 once invoked), scanned B, original, MIT.

A review agent for database, infrastructure, and CI security issues that automated grading cannot fully resolve. It examines SQL functions, row-level security policies, dynamic table references, pipelines, and container or infrastructure files.

In plain words
What is it for?
Reviewing uncertain database permissions, ownership checks, RLS policies, infrastructure configuration, CI workflows, and container security.
Why use it?
A scan may prove that security-related code exists without proving that it enforces the intended restriction.

Agent for Claude Code

Written for Claude Code: ${CLAUDE_PLUGIN_ROOT} variable. Also seen: model in frontmatter.

Runs only inside its plugin — its command needs a path that Claude Code sets for a plugin’s own hooks and for nothing else. Install the plugin, not this.

Part of the claudeguard-il plugin — 7 skills, 6 agents shipped together

Good fit Reviewing uncertain database permissions, ownership checks, RLS policies, infrastructure configuration, CI workflows, and container security.

Compare 6 agents from other repositories ↓
Install

Getting it into your agent

This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.

Claude Code
/plugin marketplace add Freespirits/claudeguard-il
Claude Code
/plugin install claudeguard-il

Made for: Claude Code.

Or install claudeguard-il, the plugin that ships this one along with the rest of its 7 skills, 6 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.

agentmods badge for infra-auditor

README.md
[![agentmods](https://agentmods.dev/badge/agents/freespirits/claudeguard-il/infra-auditor.svg)](https://agentmods.dev/agents/freespirits/claudeguard-il/infra-auditor)
Your own site
<a href="https://agentmods.dev/agents/freespirits/claudeguard-il/infra-auditor"><img src="https://agentmods.dev/badge/agents/freespirits/claudeguard-il/infra-auditor.svg" alt="Measured on agentmods" height="20"></a>
Per session 185 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 4,032 The whole file, excluding the scripts and references it only reads on demand.
Security scan B 2 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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00185 $0.04032
Opus 5 $0.00093 $0.02016
Sonnet 5 $0.00037 $0.00806
Haiku 4.5 $0.00018 $0.00403

Measured 6d ago against content hash 1c9e661fc8cd, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-07, from the pricing page.

Security

Grade B, and why

infra-auditor scanned grade B with 2 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 6d 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.

Downloads and executes remote codemediumSupply chain

curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.

(CG-IAC-002), `curl | sh` at build time (CG-IAC-003), unpinned base images (CG-IAC-004), the Docker

Downgraded: this mod is about security review, or the phrase is quoted, so it is likely naming the pattern rather than instructing it.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

(CG-IAC-002), `curl | sh` at build time (CG-IAC-003), unpinned base images (CG-IAC-004), the Docker
plugin/agents/infra-auditor.md · 239 lines

How it starts

The opening of the file, as written. The whole thing — 239 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are the database / infrastructure / CI reviewer for ClaudeGuardIL.

The grader can prove that a SQL function is SECURITY DEFINER, that it pins search_path, and that its body contains the characters auth.uid(). It cannot prove that the value is used to gate anything. A function that reads auth.uid() into a variable it never compares is indistinguishable from one that enforces ownership — from a regex. Settling that is your job, and the difference is a complete RLS bypass reachable through supabase.rpc() by anyone.

When to invoke

  • After the grader runs with rows in coverage.sqlFunctions.undeterminable, coverage.tables.undeterminable, coverage.dynamicTableRefs, or coverage.ungradedSurfaces.
  • Targeted review of RLS policies, a database function, a pipeline, or a container image.

Presence of a workflow, Dockerfile or *.tf is no longer on its own a reason to invoke: those are graded, and their fail rows are already findings. Invoke for their pass rows only when someone wants the non-mechanical review described in item 5.

Your work list

Your input is the grader's JSON plus model.artifacts. Walk these, in this order:

1. coverage.sqlFunctions.undeterminable — highest value on your list. The note reads "security definer that pins search_path and references auth.uid() — whether the check actually gates the body is not verified". SECURITY DEFINER runs with the owner's rights and bypasses RLS entirely, and anyone can invoke it through supabase.rpc(). So the function itself is the only authorization there is. Read the whole body and look for these, all of which pass the rule:

  • auth.uid() used only to populate a column (insert into t(user_id) values (auth.uid())) while the where clause of an update or delete is unfiltered. Writes are attributed correctly; nothing is gated.
  • The function takes a caller-supplied p_user_id argument and filters on that instead of auth.uid(). The caller names their own identity.
  • if auth.uid() is null then raise notice ... — a log, not a raise exception. Execution continues.
  • A guard clause whose else branch still reaches the mutating statement, or a check placed after the update/delete it was meant to protect.
  • where user_id = auth.uid() or <something permissive>, or coalesce(auth.uid(), <fallback uuid>).
  • The real work delegated to a second function called with no filter — the outer function checks, the inner one does not, and the inner one is also callable directly.

Read the full file on GitHub · 239 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.

  1. 6d ago First seen · 239 lines · 185 tokens per session scan B 1c9e661fc8cd

Subscribe to this mod's changes

infra-auditor is an agent published in the GitHub repository Freespirits/claudeguard-il (2 stars, last pushed 1mo ago), licensed MIT. It adds 185 tokens to every session and 4,032 once invoked, about $0.0009 per session on Opus 5. A static security scan graded it B with 2 findings (downloads and executes remote code, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other agents, from other repositories

MS-SQL Database Administrator

Work with Microsoft SQL Server databases using the MS SQL extension.

github/awesome-copilot · 18 tokens

core-data-auditor

Use this agent when the user mentions Core Data review, schema migration, production crashes, or data safety checking. Automatically scans Core Data code for the 5 most critical safety violations - schema migration risks, thread-confinement errors, N+1 query patterns, production data loss risks, and performance issues…

CharlesWiltgen/Axiom · 261 tokens

lens

Turns raw data into actionable decisions — dashboards, metric definitions, SQL analytics, funnel and cohort analysis across BI platforms. Use when designing a dashboard, defining KPIs, or running funnel analysis. Trigger with "design a dashboard", "analyze our funnel".

jeremylongshore/tons-of-skills-marketplace · 53 tokens

ecto-schema-designer

Ecto schema architect - designs migrations, data models, and query patterns. Use proactively when planning database structure for new features.

oliver-kriska/claude-elixir-phoenix · 30 tokens

django-migrations-specialist

Database specialist for Django, runs in the "database" extra phase after development. Finalizes model field types and Meta indexes/constraints, runs makemigrations, reviews generated SQL with sqlmigrate, runs migrate, verifies with migrate --check. Do NOT use for: application logic (django-architect), tests…

AratKruglik/claude-sdlc · 85 tokens

sql-expert

Usa este agente para cualquier tarea relacionada con base de datos en FacturaScripts: diseñar esquemas de tabla XML, optimizar consultas con DbQuery y Where, crear índices y constraints, escribir migraciones SQL, analizar rendimiento de queries, usar transacciones, trabajar con DataBaseWhere/DataBase/DbQuery, diseñar…

FacturaScripts/fs-claude-plugin · 102 tokens