security-audit

A read-only review of the security settings and findings in AWS and GitHub. It examines the running cloud and code-hosting setup, rather than reviewing a particular code change.

In plain words
What is it for?
Use it to collect security findings, review data-protection and monitoring controls, and create a prioritized remediation plan.
Why use it?
It helps identify infrastructure security gaps without making changes to accounts, resources, or repositories.

Command for Claude Code

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.

agentmods
npx agentmods add commands/robofinsystems/robosystems/security-audit
Clone the repo
git clone --depth 1 https://github.com/RoboFinSystems/robosystems

Made for: Claude Code.

Per session 16 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 3,606 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00016 $0.03606
Opus 5 $0.00008 $0.01803
Sonnet 5 $0.00003 $0.00721
Haiku 4.5 $0.00002 $0.00361

Measured 2d ago against content hash 56d1e706a6a5, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

security-audit 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 2d 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.

.claude/commands/security-audit.md · 162 lines

How it starts

The opening of the file, as written. The whole thing — 162 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Run a read-only AWS + GitHub security-posture review, then produce a findings report and a prioritized remediation plan.

Scope boundary. This is an infrastructure/cloud posture audit — detective controls, live findings, data protection. It is not a code review. /security-review reviews pending code changes on the current branch; /pr-review covers a PR's diff. Those look at what the code does; this looks at what the account is running. Keep them separate — don't fold diff review into this sweep.

Pairs with the security-posture-audit runbook in local/RoboSystems/runbooks/read it alongside this file. This file is the public, generic methodology; the runbook is the live, account-specific record of what's currently enabled, suppressed, accepted, or retired, and it is authoritative for this environment.

Scope & guardrails

  • Read-only by default. Use only describe-* / list-* / get-* and gh reads. Never run get-secret-value. Any change — enabling a service, creating a suppression/archive rule, deleting a resource, deploying a stack — requires explicit in-the-moment user confirmation, and destructive/CloudFormation actions are the user's to run. One carve-out: generate-credential-report is technically a write (it produces a report) but carries no configuration change — it's in scope for Phase 4.
  • Outputs are sensitive — never commit them. This skill (the methodology) is safe in a public repo; its output — a report of live findings, exposed resources, and disabled controls — is a reconnaissance roadmap for an attacker. Write reports to a git-ignored path (local/, the scratchpad) or an ephemeral private Artifact. Never commit a findings report, account ID, resource ARN, or CVE inventory to the repo.
  • Never state live posture in this file. Naming the configuration — a stack, a log group, an ECR repo, a GitHub variable — is fine and makes the sweep usable; it's already in the public IaC. Naming its current value is not: no "X is enabled", no finding counts, no account IDs or ARNs. Those rot silently and are a disclosure besides. Every posture question below resolves to a command you run now; anything you learn from running it belongs in the runbook, not here.
  • Context. Discover the account at runtime (aws sts get-caller-identity); region defaults to us-east-1; the robosystems deployment runs prod + staging in one shared account, so expect per-environment resource pairs. Detective services are account-global singletons gated by GitHub variables (see Phase 6). Frontend images live in ECR repos robosystems-app / roboledger-app / roboinvestor-app; the backend image is robosystems. Confirm what's actually deployed with aws cloudformation list-stacks --stack-status-filter CREATE_COMPLETE UPDATE_COMPLETE.
  • Parallelize the sweep. Phases 1–6 are independent read-only sweeps — fan them out with the Agent tool (one agent per phase) and synthesize the results in the main loop. A denied/empty CLI call is itself a finding (service not enabled, or the role lacks the read perm — say which).

Read the full file on GitHub · 162 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. 2d ago First seen · 162 lines · 16 tokens per session scan A 56d1e706a6a5

Subscribe to this mod's changes

security-audit is a command published in the GitHub repository RoboFinSystems/robosystems (24 stars, last pushed 3d ago), licensed Apache-2.0. It adds 16 tokens to every session and 3,606 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 2026-08-30.