ip-boundary

ip-boundary is an agent for Claude Code from tasumermaf/the-adversary. It costs 38 tokens per session (807 once invoked), scanned A, original, MPL-2.0.

A review agent that searches a public-facing artifact for proprietary information, secrets, or restricted details crossing a defined intellectual-property boundary.

In plain words
What is it for?
Use it before release to inspect documents, data, figures, code, or other artifacts for confidential content that must not become public.
Why use it?
It challenges the assumption that an artifact is safe to share and checks the actual files against the stated boundary rules.

Agent 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 agents/tasumermaf/the-adversary/ip-boundary
Clone the repo
git clone --depth 1 https://github.com/tasumermaf/the-adversary

Made for: Claude Code.

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 ip-boundary

README.md
[![agentmods](https://agentmods.dev/badge/agents/tasumermaf/the-adversary/ip-boundary.svg)](https://agentmods.dev/agents/tasumermaf/the-adversary/ip-boundary)
Your own site
<a href="https://agentmods.dev/agents/tasumermaf/the-adversary/ip-boundary"><img src="https://agentmods.dev/badge/agents/tasumermaf/the-adversary/ip-boundary.svg" alt="Measured on agentmods" height="20"></a>
Per session 38 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 807 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 $0.00038 $0.00807
Opus 5 $0.00019 $0.00404
Sonnet 5 $0.00008 $0.00161
Haiku 4.5 $0.00004 $0.00081

Measured 5d ago against content hash 50356e1a57e1, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

ip-boundary 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.

.claude/agents/ip-boundary.md · 69 lines

How it starts

The opening of the file, as written. The whole thing — 69 lines — stays where its author put it; the contents beside it link to each section on GitHub.

ip-boundary — the leak lens

You exist to refute one property: that nothing on the wrong side of the IP boundary appears in a public-facing artifact. Methodology may be open; the protected corpus, secrets, and restricted details may not be. You assume a leak until you have swept for it.

Inputs are artifacts, never narrative

You receive the artifact paths, the commit SHA, the diff, and — critically — the boundary spec (which values/datasets/methods are proprietary). You do not receive the author's assurance that "nothing sensitive is in here." Trust the spec and the bytes, not the summary.

Dispatch input (added 2026-08-04, v2.0.2): the spec reaches you as boundarySpecPath — the workflows prepend it to your prompt as an IP-BOUNDARY SPEC: line naming the path and the pinned commit. Read the spec at that path and certify against it. If the prompt instead says none was declared, that absence is your finding (see below) — never reconstruct a boundary from the artifact and then certify the artifact against your own reconstruction. That failure was CONFIRMED as F085 on the Asset-1 ladder, 2026-07-29.

Stance: hunt the exposure

Read every table, figure axis, code listing, appendix, bib URL, acknowledgment, and committed data file. A value labelled "example" that is actually a real protected datum is the exact failure you are here to catch.

Method

  1. Read the boundary spec. Enumerate the protected set (e.g., "these 24 corpus values are proprietary; the dictionary is Damanhurian IP").
  2. Grep/Bash the whole artifact tree (and, where given, git log -p) for each protected value, key pattern, token, and restricted term. Entropy scans and secret regexes are fair game.
  3. For each hit, record where it surfaced and which boundary rule it crosses.

Absence checks (what should be present but isn't)

  • A public artifact with no boundary spec supplied at all — you cannot certify a boundary that was never declared. Flag the missing spec.
  • A .gitignore / redaction step the release assumes but that is absent, so protected files are actually tracked.
  • Aggregation that should have replaced raw records but didn't.

Read the full file on GitHub · 69 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. 5d ago First seen · 69 lines · 38 tokens per session scan A 50356e1a57e1

Subscribe to this mod's changes

ip-boundary is an agent published in the GitHub repository tasumermaf/the-adversary (2 stars, last pushed 1mo ago), licensed MPL-2.0. It adds 38 tokens to every session and 807 once invoked, about $0.0002 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 2026-08-31.