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.
/plugin marketplace add cognitive-fab/polygraphnpx agentmods add plugins/cognitive-fab/polygraph/polyvizgit clone --depth 1 https://github.com/cognitive-fab/polygraphGrade A, and why
polyviz 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.
What it actually says
{
"name": "polyviz",
"displayName": "polyviz",
"version": "0.1.0",
"description": "A deterministic, artifact-derived diagram renderer for Polygraph. Turns the artifacts the suite already produces — the state machine, the invariants, the counterexample, the polyvers compat verdict — into a fixed catalog of clean, brand-consistent SVGs (optional PNG). Same inputs → byte-identical output; no model call at render time. Renders from a viz-model JSON or straight from a Polygraph/polyvers artifacts directory (adapters derive the machine graph by bounded BFS over the module, read the invariants, resolve a failing findings.json to its counterexample trace, and map compat-report.json). Optional runtime components: the state-machine graph needs elkjs and PNG needs @resvg/resvg-js (both optionalDependencies, loaded lazily); the pure viz-model → SVG path needs neither and executes no user code. Deriving the machine graph from a real run does execute the target module (bounded reachability).",
"author": {
"name": "jdubray",
"url": "https://github.com/jdubray"
},
"homepage": "https://github.com/cognitive-fab/polygraph",
"repository": "https://github.com/cognitive-fab/polygraph",
"license": "Apache-2.0",
"keywords": [
"visualization",
"diagram",
"svg",
"state-machine",
"verification",
"polygraph",
"polyvers",
"deterministic",
"counterexample"
]
}
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.
- yesterday First seen · 25 lines scan A 62dd7fa32e0f
polyviz is a plugin published in the GitHub repository cognitive-fab/polygraph (11 stars, last pushed 5d ago), licensed Apache-2.0. Its token cost is not measured: this kind of file is read by the harness, not the model. 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.
Other plugins, from other repositories
intlayer
Intlayer plugins for Claude Code — LSP-powered i18n intelligence for TypeScript, JavaScript, Vue, and Svelte projects.
fablecut
FableCut — a browser video editor that AI agents can drive.
autonomous-dev
PROJECT.md-first autonomous development with 22 specialized agents, 28 skills, and 66 automation hooks. Full SDLC automation via /implement command. Zero scope creep with automatic PROJECT.md validation. Works for Python, JavaScript, TypeScript, and Go projects.
js-reverse-plugin
JavaScript AST deobfuscation, controlled iv8 execution, JSVMP recovery, and browser evidence through js-reverse-mcp.
ai-reverse-resources
AI-assisted reverse engineering plugins for Codex and Claude Code.
repo-scaffold
Analyzes your repository and generates CLAUDE.md and project-specific skills.