state-graph-mcp

A set of rules for tracking project tasks, decisions, dependencies, and blockers in a state graph. A state graph is a linked record of what work exists, what depends on it, and what is preventing progress.

In plain words
What is it for?
Use it to review project status, find blockers and pending tasks, trace dependencies, record decisions, create artifact records, and mark completed work.
Why use it?
It gives the agent a current project map before coding and keeps important work status and decisions recorded as the task progresses.

Cursor rule for Cursor

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 rules/putervision/webcrypt/state-graph-mcp
Clone the repo
git clone --depth 1 https://github.com/putervision/WebCrypt

Made for: Cursor.

Per session 562 This file is loaded in full into every session.
When invoked 562 The same file — it is already loaded in full.
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 $0.00562 $0.00562
Opus 5 $0.00281 $0.00281
Sonnet 5 $0.00112 $0.00112
Haiku 4.5 $0.00056 $0.00056

Measured yesterday against content hash 74394a76b86d, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

state-graph-mcp 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 yesterday.

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.

.cursor/rules/state-graph-mcp.mdc · 31 lines

How it starts

The opening of the file, as written. The whole thing — 31 lines — stays where its author put it; the contents beside it link to each section on GitHub.

State Graph (state-graph-mcp)

This project tracks workflow state, tasks, design decisions, and blockers using state-graph-mcp.

1. Priority Order

Before doing any coding or investigation:

  1. get_project_summary — Always run this first to understand current project state, active branches, and overall progress.
  2. find_blockers — Identify any active blockers preventing progress.
  3. list_nodes — Find pending tasks, past decisions, or milestones.
  4. trace_dependencies — Trace what depends on or blocks a task.

2. When to Write to the Graph

You MUST update the graph as you work:

  • Starting a new task: Create a node with add_node(type: "task", title: "...").
  • Making a design or implementation decision: Document it with add_node(type: "decision", title: "...", metadata: { "rationale": "..." }).
  • Encountering a blocker: Record the blocker with add_node(type: "blocker", ...) and connect it to the blocked task using add_edge(type: "blocks", source_id: blocker_id, target_id: task_id).
  • Completing a task: Update the status to done using update_node(id: task_id, status: "done").
  • Creating/generating a new file: Create an artifact node with add_node(type: "artifact", ...) and connect it using add_edge(type: "produces").

3. Workflow Pattern

  1. Start of session: Run get_project_summary and find_blockers to align on current status.
  2. Task decomposition: Decompose user requests into tasks and add them to the graph.
  3. Execution: Mark tasks as "in_progress", document design decisions as they occur, and log blockers if you hit any obstacles.
  4. Resolution: Mark tasks as "done", document completed artifacts, and resolve blockers.

4. Codebase Seeding on Initialization

If the project was just initialized or is missing high-level structure (Plans, Milestones, Decisions):

  1. Inspect the Codebase: Read the README and core files to understand the roadmap and architecture.
  2. Scaffold the Roadmap: Create a plan node (e.g., "Project Roadmap") and add milestone nodes representing key target phases, connecting them using part_of edges.
  3. Scaffold Architecture: Create decision nodes representing core technical choices (e.g., choice of databases, frameworks) and link them to the milestones/tasks using decided_in edges.

Read the full file on GitHub · 31 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. yesterday First seen · 31 lines · 562 tokens per session scan A 74394a76b86d

Subscribe to this mod's changes

state-graph-mcp is a cursor rule published in the GitHub repository putervision/WebCrypt (26 stars, last pushed 10d ago), licensed MIT. It adds 562 tokens to every session, about $0.0028 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 2026-08-30.