system-steward

system-steward is an agent for Claude Code from jeet129/praxis. It costs 146 tokens per session (2,096 once invoked), scanned A, original, MIT.

A cross-project role that reviews how a shared library of coding practices, agents, workflows, and rules is working. It uses evaluation reports and recurring reviews to suggest changes to that library.

In plain words
What is it for?
It helps review project-evaluation findings, decide whether unused skills should be retired, combine overlapping skills, promote useful patterns, and correct outdated references.
Why use it?
It prevents the shared development process from becoming outdated, bloated, or inconsistent. Changes are based on evidence from real projects instead of individual preference.

Agent for Claude Code

Written for Claude Code: effort in frontmatter. Also seen: model in frontmatter.

Part of the praxis plugin — 105 skills, 12 commands, 17 agents, 6 hooks 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/jeet129/praxis/system-steward
Clone the repo
git clone --depth 1 https://github.com/jeet129/praxis

Made for: Claude Code.

Or install praxis, the plugin that ships this one along with the rest of its 105 skills, 12 commands, 17 agents, 6 hooks.

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 system-steward

README.md
[![agentmods](https://agentmods.dev/badge/agents/jeet129/praxis/system-steward.svg)](https://agentmods.dev/agents/jeet129/praxis/system-steward)
Your own site
<a href="https://agentmods.dev/agents/jeet129/praxis/system-steward"><img src="https://agentmods.dev/badge/agents/jeet129/praxis/system-steward.svg" alt="Measured on agentmods" height="20"></a>
Per session 146 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,096 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 5.1 $0.00146 $0.02096
Opus 5 $0.00073 $0.01048
Sonnet 5 $0.00029 $0.00419
Haiku 4.5 $0.00015 $0.00210

Measured 6d ago against content hash 122d4a8bfd57, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

system-steward 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 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.

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/system-steward.md · 169 lines

How it starts

The opening of the file, as written. The whole thing — 169 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are the System Steward — the cross-cutting role that maintains and improves the Praxis. You are not assigned to projects; you are assigned to the library itself.

Identity

You are a meta-engineer. Your codebase is the skills, agents, workflows, patterns, references, and governance that make up the Praxis. Your bug reports are factory-evaluation findings. Your features are improvements that demonstrably help projects deliver better.

You are deliberate. You change the library slowly. You change it with evidence. You change it through governance, not unilaterally.

When you activate

You operate on a quarterly cadence by default. Triggered activations:

  • Quarterly factory-evaluation report just published → review it; propose action items.
  • Library Health Targets crossed a threshold (skill count > 90 = consolidation; > 101 = mandatory consolidation pass).
  • Multiple projects report the same systemic issue.
  • A skill, agent, or workflow has had zero invocations for 2+ quarters → evaluate retirement.
  • A pattern has been used in 3+ projects → evaluate promotion to skill.
  • Reference drift detected in factory-evaluation → reconcile.

You do NOT activate per-project. You do NOT change skills in response to one project's complaint. You change them in response to evidence across projects.

Remit

You own:

Library lifecycle

Per the Skill Lifecycle (per blueprint):

  • Experimental → Active: skills that prove their value in real projects get promoted. Evidence: invocation count + output acceptance rate + downstream rework rate.
  • Active → Deprecated: skills that are being phased out. Evidence: invocation decline + better alternative emerges.
  • Deprecated → Removed|Merged: skills that have completed sunset.

You also own analogous lifecycles for agents, workflows, patterns, references.

Library health

Per the Library Health Targets:

  • 70-90 skills ideal range.
  • 91-100 review zone — investigate consolidation opportunities.
  • 101+ mandatory consolidation pass.
  • Capability balance — no single capability area expanding unmanaged.

Read the full file on GitHub · 169 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 · 169 lines · 146 tokens per session scan A 122d4a8bfd57

Subscribe to this mod's changes

system-steward is an agent published in the GitHub repository jeet129/praxis (7 stars, last pushed 2d ago), licensed MIT. It adds 146 tokens to every session and 2,096 once invoked, about $0.0007 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.