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.
git clone --depth 1 https://github.com/gsaranti/pantheonWrote 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.
[](https://agentmods.dev/plugins/gsaranti/pantheon/claude-code)<a href="https://agentmods.dev/plugins/gsaranti/pantheon/claude-code"><img src="https://agentmods.dev/badge/plugins/gsaranti/pantheon/claude-code/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.
<a href="https://agentmods.dev/plugins/gsaranti/pantheon/claude-code"><img src="https://agentmods.dev/badge/plugins/gsaranti/pantheon/claude-code.svg" alt="Reviewed on agentmods" width="80" height="20"></a>Grade A, and why
iris 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 1mo 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.
What it actually says
{
"$schema": "https://json.schemastore.org/claude-code-plugin-manifest.json",
"name": "iris",
"displayName": "Iris",
"version": "0.2.0",
"description": "Cross-tool context bridge between Claude Code and Codex. Logs your conversation and exposes it to the other tool so you can hand off plans, share context, and pick up where the other tool left off.",
"license": "MIT",
"author": {
"name": "George Sarantinos"
},
"homepage": "https://github.com/gsaranti/pantheon/tree/main/plugins/iris",
"repository": "https://github.com/gsaranti/pantheon/tree/main/plugins/iris",
"keywords": ["pantheon", "claude-code", "codex", "context", "handoff", "interop", "messenger", "bridge", "cross-tool"]
}
What it installs
The manifest is a name and a version. 4 skills, 3 hooks travel with it, and installing the plugin installs all of them — 202 tokens a session between them. Each is measured on its own page, and each can be installed alone.
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.
- 1mo ago First seen · 15 lines scan A 57bcd69818d6
iris is a plugin published in the GitHub repository gsaranti/pantheon (3 stars, last pushed 1mo ago), licensed MIT. 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
alive
Local, user-owned context for Claude Code. Keep your work in files you control.
evolving-lite
Self-evolving Claude Code system that learns from corrections, manages context, and improves every session.
zc-ctx
Secure memory + persistent context + skill-admission gate for Claude Code. HMAC-chained toolcalls + outcomes, per-agent HKDF subkey isolation, Postgres RLS, work-stealing task queue, MemGPT-style persistent memory, Agent Notebook Model (Read→Summary redirect), and v0.26.0+ Anthropic-style filesystem skills with AST.
furl
Furl context compression for Claude Code: bundles the Furl MCP server (furlcompress / furlretrieve / furlstats / furlpurge / furlsearch / furllist) and a fail-open PostToolUse hook that shrinks large tool outputs before they land in context. Originals stay byte-exactly retrievable via > markers.
consolidate-memory
Cross-project, verification-first memory for Claude Code — the layer beyond the built-in Auto Dream. Shares memory across enrolled projects in the same trust domain (a governed store, replicated into those projects) and verifies facts against the live code (grep/file/git), dropping what it can't confirm — fact-checked.
dream-beta-tester
Beta-tests the consolidate-memory "dream" skill ITSELF — the QA companion to it. A deterministic invariant oracle plus a judgment-lens skill (/dream-beta-test) that finds defects, regressions, and dishonest reports in the dream/consolidation skill, honest by construction (confirmed defects are kept separate from…