Home/Catalog/Reliability
Reliability — will it stay up?
Brief 124

Abuse & Overload Protection

Whether the system defends itself against traffic that is not a well-behaved user — bots, scrapers, floods, and expensive requests that can exhaust it.

ABUSE.md4 phases7 lenses~2.8k chars

Free & open · no signup · read-only — it ends by asking · nothing leaves your machine

What it does

You are working inside this repo. Mission: judge how the system holds up against hostile or careless load — the automated abuse, the floods, and the disproportionately expensive requests that can exhaust resources or run up a bill. This is defense against overload, distinct from designing fair API limits.

Read-only pass. Read the exposed endpoints, their cost, and any protections; change nothing but the report file.

What it looks for

7 lenses it looks through

Phase 2 sweeps the codebase through every one of these, citing file and line for each finding.

1

Expensive endpoints

requests that cost far more than they charge or check

2

Unauthenticated exposure

costly operations reachable with no login and no cost

3

Bot & scrape resistance

defenses against automated abuse, enumeration, and harvesting

4

Overload behavior

load shedding and prioritization under flood, or everything falling together

5

Amplification

one request that triggers many: fan-out, recursion, unbounded work

6

Resource exhaustion

memory, connections, and disk an attacker or bug can consume

7

Abuse detection

would sustained abuse be noticed and throttled, or run until it bills or breaks

The deliverable

What lands in your repo

One structured report at the repo root — or in reports/, if you keep one — the same shape every time, ready for a teammate — or the optional Studio — to act on.

ABUSE.md
1
Exposed & expensive
the surfaces most attractive to abuse, with their cost
2
Findings
each: risk · surface · ease and cost · the defense
3
Overload behavior
what happens under flood today, and the shedding to add
4
Priority
the defenses to add first, by ease of exploitation and blast radius
One file. Evidence-backed. It ends by asking before touching anything.
How it works

What it makes the agent do

Every Goal Prompt follows the same four steps, so results are consistent and repeatable — no matter which one you run.

1
Phase 1

Find the expensive and exposed

List the operations that cost far more than a normal request.

2
Phase 2

Audit through 7 lenses

Expensive endpoints — requests that cost far more than they charge or check.

3
Phase 3

Curate

Rank by ease of exploitation × cost of the damage.

4
Phase 4

Report

Create ABUSE.md at repo root.

Get started

Three ways to run this Goal Prompt

01 · COPY

Paste it in

Copy it and paste it into your agent inside the repo you want checked.

02 · INSTALL

As a slash command

Install the goal plugin once — two commands — then just type /goal:abuse-overload.

1/plugin marketplace add GhostlyGawd/goal-prompts
2/plugin install goal@goal-prompts

Or install only this Goal Prompt as /goal-abuse-overload:

curl -fsSL https://goal-prompts.vercel.app/install | BRIEF=124 sh
03 · AGENT

From an agent (MCP)

Let an agent fetch it mid-conversation, or pull the raw Goal Prompt by URL.

https://goal-prompts.vercel.app/raw/124.md
Transparency

The exact prompt

Nothing hidden — this is the whole Goal Prompt, verbatim. Read it in a minute, edit it, or copy it as-is.

Read the full brief (2,769 characters)
# Goal: Abuse & Overload Protection

You are working inside this repo. Mission: judge how the system holds up against hostile or careless load — the automated abuse, the floods, and the disproportionately expensive requests that can exhaust resources or run up a bill. This is defense against overload, distinct from designing fair API limits.

Read-only pass. Read the exposed endpoints, their cost, and any protections; change nothing but the report file.

## Phase 1 — Find the expensive and exposed
- List the operations that cost far more than a normal request: heavy queries, exports, AI calls, fan-outs.
- Note which of those are reachable without authentication or payment.
- See what protects the system when load spikes.

## Phase 2 — Audit through 7 lenses
1. **Expensive endpoints** — requests that cost far more than they charge or check
2. **Unauthenticated exposure** — costly operations reachable with no login and no cost
3. **Bot & scrape resistance** — defenses against automated abuse, enumeration, and harvesting
4. **Overload behavior** — load shedding and prioritization under flood, or everything falling together
5. **Amplification** — one request that triggers many: fan-out, recursion, unbounded work
6. **Resource exhaustion** — memory, connections, and disk an attacker or bug can consume
7. **Abuse detection** — would sustained abuse be noticed and throttled, or run until it bills or breaks

## Phase 3 — Curate
- Rank by ease of exploitation × cost of the damage: an unauthenticated expensive endpoint tops the list.
- For each, name the defense — auth, a cost ceiling, load shedding, bot mitigation, detection.
- Separate "runs up a bill" from "takes the system down"; both are outages of a kind.

## Phase 4 — Report
Create `ABUSE.md` at repo root:
1. **Exposed & expensive** — the surfaces most attractive to abuse, with their cost
2. **Findings** — each: risk · surface · ease and cost · the defense
3. **Overload behavior** — what happens under flood today, and the shedding to add
4. **Priority** — the defenses to add first, by ease of exploitation and blast radius

Start the report with today's date. If `ABUSE.md` already exists from a previous run, read it first and lead with what changed since.

## Rules
- An unauthenticated expensive endpoint is a denial-of-wallet waiting to happen
- Under overload, shed load deliberately; do not let everything fail together
- No public-facing endpoints in this repo? Say so in a one-paragraph null report and stop — a null result is a valid finding.
- If a `reports/` directory exists at the repo root, write the report there instead of the root.
- Before asking, present the top findings as a ranked list in plain words
- Report only — end by asking which abuse defenses to add first

House rules for this Goal Prompt

  • An unauthenticated expensive endpoint is a denial-of-wallet waiting to happen
  • Under overload, shed load deliberately; do not let everything fail together
  • No public-facing endpoints in this repo? Say so in a one-paragraph null report and stop — a null result is a valid finding.
  • If a reports/ directory exists at the repo root, write the report there instead of the root.
  • Before asking, present the top findings as a ranked list in plain words
  • Report only — end by asking which abuse defenses to add first
Keep exploring

More Reliability briefs

Runs inside these playbooks — curated sequences you can launch with one paste: