superpowers-graph-monitor

superpowers-graph-monitor is an agent for Claude Code from RonMizrahi/superpowers-graph. It costs 113 tokens per session (1,796 once invoked), scanned A, original, MIT.

A read-only checker for runs managed as a graph of steps and rules. It compares the recorded run with what the graph says should have happened.

In plain words
What is it for?
Use it to audit a graph run using its state file, graph specification, repository, and optionally a transcript or software-development ledger. A transcript is needed for transcript-only checks.
Why use it?
It can reveal skipped steps, incorrect retry counts, unsafe ordering, unmatched exit conditions, and decisions that appear only in the conversation record.

Agent for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: mentions subagents.

Part of the superpowers-graph plugin — 1 skill, 1 agent shipped together

Good fit Use it to audit a graph run using its state file, graph specification, repository, and optionally a transcript or software-development ledger. A transcript is needed for transcript-only checks.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/ronmizrahi/superpowers-graph/superpowers-graph-monitor
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.

Clone the repo
git clone --depth 1 https://github.com/RonMizrahi/superpowers-graph

Made for: Claude Code.

Or install superpowers-graph, the plugin that ships this one along with the rest of its 1 skill, 1 agent.

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 superpowers-graph-monitor

README.md
[![agentmods](https://agentmods.dev/badge/agents/ronmizrahi/superpowers-graph/superpowers-graph-monitor/github.svg)](https://agentmods.dev/agents/ronmizrahi/superpowers-graph/superpowers-graph-monitor)
Your own site
<a href="https://agentmods.dev/agents/ronmizrahi/superpowers-graph/superpowers-graph-monitor"><img src="https://agentmods.dev/badge/agents/ronmizrahi/superpowers-graph/superpowers-graph-monitor/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for superpowers-graph-monitor

Your own site · 80×15
<a href="https://agentmods.dev/agents/ronmizrahi/superpowers-graph/superpowers-graph-monitor"><img src="https://agentmods.dev/badge/agents/ronmizrahi/superpowers-graph/superpowers-graph-monitor.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 113 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,796 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 $0.00113 $0.01796
Opus 5 $0.00056 $0.00898
Sonnet 5 $0.00023 $0.00359
Haiku 4.5 $0.00011 $0.00180

Measured 12d ago against content hash 92a68943821e, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-12, from the pricing page.

Security

Grade A, and why

superpowers-graph-monitor 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 12d 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.

plugins/superpowers-graph/agents/superpowers-graph-monitor.md · 142 lines

How it starts

The opening of the file, as written. The whole thing — 142 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Superpowers Graph Monitor

Goal

One question: is this run doing what the graph says it is doing?

Not was the work any good — that is a human question. Only whether the record and the world agree with the model.

Inputs

state_path — the run-state JSON · graph_root — the skill directory, read as the spec (nodes.md, edges.md, state.md) · repo — the repository the run operated on · sdd_ledger (optional) · transcript (optional — without it the transcript-only checks cannot run, and you must say so).

Checks

0. Run the graph's own evals first — they answer most of the structural questions in a second

python3 <graph_root>/evals/run_all.py

spec_consistency.py covers vocabulary, published counts, cycle counts, bound agreement, the DEBUG caller list and whether the ordering invariant is still stated in a checkable form. graph_walk.py replays scripted runs against edges.md and state.md and fails if the fixture set no longer reaches every node and every edge. A red suite is the strongest possible answer to "is this graph sound?" and you have it before reading a line of the run.

A green suite is not a clean bill of health. It checks the spec and the fixtures, not this run — everything below is still yours to do. If the suite is absent or will not run, say so and fall back to checking by hand; never report an unrun check as a pass.

1. Trail integrity

  • Every history[].guard appears verbatim in edges.md.
  • Every history[].edge id exists in edges.md, and its from/to match that row.
  • history[i].to == history[i+1].from for every consecutive pair.
  • The last entry's to equals state.node.
  • No history[].edge id appears in the Retired edge ids section.

2. Zero-guard halts, classified correctly

blocked.guards_tested.length > 1 && blocked.tried == [] means a zero-match halt — the node returned a legal result no guard covered. blocked.tried non-empty with a single guard means bound exhaustion. These need completely different fixes; a report that conflates them sends someone to widen a bound when the defect is a missing guard.

Read the full file on GitHub · 142 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. 12d ago First seen · 142 lines · 113 tokens per session scan A 92a68943821e

Subscribe to this mod's changes

superpowers-graph-monitor is an agent published in the GitHub repository RonMizrahi/superpowers-graph (1 stars, last pushed 1mo ago), licensed MIT. It adds 113 tokens to every session and 1,796 once invoked, about $0.0006 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.

Related

Other agents, from other repositories

queue-debugger

Use this agent when the user encounters "queue not delivering messages", "consumer errors", "DLQ filling up", "throughput issues", "message backlog", "queue timeout errors", or needs systematic Cloudflare Queues troubleshooting. Examples.

secondsky/claude-skills · 0 tokens

d1-debugger

Autonomous diagnostic agent that investigates Cloudflare D1 database issues through 9-phase analysis (config, migrations, queries, bindings, errors, limits, performance, Time Travel, report). Use when encountering D1 query errors, migration failures, binding issues, performance degradation, or limit/quota errors.

secondsky/claude-skills · 65 tokens

do-debugger

Autonomous Durable Objects debugger. Automatically detects and fixes DO configuration errors, runtime issues, and common mistakes without user intervention.

secondsky/claude-skills · 28 tokens

Turnstile CSP Debugger

This agent should be used when the user asks to "validate CSP for turnstile", "fix CSP errors", "check content security policy", or encounters error 200500. Analyzes Content Security Policy headers and suggests Turnstile-compatible configurations.

secondsky/claude-skills · 58 tokens

workflow-debugger

Autonomous Cloudflare Workflows debugger. Automatically detects and fixes workflow configuration errors, runtime issues, and common mistakes without user intervention.

secondsky/claude-skills · 30 tokens

troubleshooting-agent

This agent should be used when the user asks to "debug Cloudflare Images errors", "troubleshoot upload failures", "diagnose CORS issues", "fix transformation errors", or encounters errors 5408, 9401-9413, multipart/form-data encoding issues, or API connectivity problems.

secondsky/claude-skills · 66 tokens