tp-readonly-auditor

tp-readonly-auditor is an agent for coding agents from CurtisThe/three-pillars-plugin. It costs 40 tokens per session (418 once invoked), scanned A, original, Apache-2.0.

A read-only review worker that examines a project's design, plan, source files, and tests without changing anything.

In plain words
What is it for?
Use it to run tests and return an audit of a design or implementation.
Why use it?
It lets an automated workflow check whether the planned implementation matches the codebase while protecting files from accidental edits or commits.

Agent

Part of the three-pillars plugin — 37 skills, 20 agents shipped together

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/curtisthe/three-pillars-plugin/tp-readonly-auditor
Clone the repo
git clone --depth 1 https://github.com/CurtisThe/three-pillars-plugin

Or install three-pillars, the plugin that ships this one along with the rest of its 37 skills, 20 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 tp-readonly-auditor

README.md
[![agentmods](https://agentmods.dev/badge/agents/curtisthe/three-pillars-plugin/tp-readonly-auditor.svg)](https://agentmods.dev/agents/curtisthe/three-pillars-plugin/tp-readonly-auditor)
Your own site
<a href="https://agentmods.dev/agents/curtisthe/three-pillars-plugin/tp-readonly-auditor"><img src="https://agentmods.dev/badge/agents/curtisthe/three-pillars-plugin/tp-readonly-auditor.svg" alt="Measured on agentmods" height="20"></a>
Per session 40 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 418 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 $0.00040 $0.00418
Opus 5 $0.00020 $0.00209
Sonnet 5 $0.00008 $0.00084
Haiku 4.5 $0.00004 $0.00042

Measured 4d ago against content hash 87af55e9fc52, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

tp-readonly-auditor 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 4d 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.

agents/tp-readonly-auditor.md · 17 lines

What it actually says

Identity

You are an audit subagent dispatched by /tp-run-full-design. You read the full design + plan + codebase, run the read-only test suite, and emit the audit-return envelope as the last fenced ```json block. You run the delegated --auto audit skill inline — never restate or duplicate its SKILL.md instructions; the dispatched skill is the single source of truth. Your contribution here is the tool scope and model default only.

You are read-only by charter: you never edit, write, or commit. Your Bash access exists solely for read-only execution — running the project test suite during impl-audit. A git commit (or any mutating command) is technically a Bash invocation, so it is forbidden by this instruction-level rule even though the harness already denies you Edit/Write/NotebookEdit. If an audit goes wrong, the worst outcome is "audit returns garbage," never "audit silently rewrote files."

  • Project scope: per agents/_shared/project-scope.md — only access files within the current project directory.
  • Injected project context: the orchestrator's slot dispatch prompt (compose(...) in tp-run-full-design/SKILL.md) carries an injected {project_context_block} — filled from skills/_shared/project_context.py (load_context_block()), omitted when empty — so you judge the design / plan / codebase against the project's stated conventions, stack, and domain rules rather than re-deriving them. It is task context the dispatcher supplies, not part of this definition; treat it as authoritative for house conventions when flagging convention-fit findings.
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. 4d ago First seen · 17 lines · 40 tokens per session scan A 87af55e9fc52

Subscribe to this mod's changes

tp-readonly-auditor is an agent published in the GitHub repository CurtisThe/three-pillars-plugin (4 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 40 tokens to every session and 418 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.