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/Jujitae/invaraWrote 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/jujitae/invara/plugin)<a href="https://agentmods.dev/plugins/jujitae/invara/plugin"><img src="https://agentmods.dev/badge/plugins/jujitae/invara/plugin/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/jujitae/invara/plugin"><img src="https://agentmods.dev/badge/plugins/jujitae/invara/plugin.svg" alt="Reviewed on agentmods" width="80" height="20"></a>Grade A, and why
invara 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
{
"name": "invara",
"description": "Independent verification that a change kept the constraints it declared. Seal a contract before the work: the paths that must not change, and the commands that must pass. Afterwards INVARA judges from file digests and exit codes and returns BLOCK, UNVERIFIABLE, HUMAN_REVIEW or PASS, naming which rule decided. UNVERIFIABLE means a check could not be run at all, so the work is unverified and NOT accepted. There is no model in the verdict path and no field anywhere for an agent to assert that the work is done - an agent can ask for a verdict here and cannot give one. Zero dependencies; runs on the standard library.",
"version": "0.3.0",
"author": {
"name": "MIGARYOS"
},
"homepage": "https://migaryos.com/invara/",
"repository": "https://github.com/Jujitae/invara",
"license": "Apache-2.0",
"keywords": [
"verification",
"ai-agents",
"provenance",
"evidence",
"ci",
"agent-verification",
"determinism",
"hash-chain"
]
}
What it installs
The manifest is a name and a version. 3 commands, 1 MCP server travel with it, and installing the plugin installs all of them — 119 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 Changed a929b4ea2988
- 1mo ago Changed 4c3a3b397277
- 1mo ago Changed 437b0be620c4
- 1mo ago Changed d7da25a89f86
- 1mo ago First seen · 22 lines scan A 545f970cf9f5
invara is a plugin published in the GitHub repository Jujitae/invara (1 stars, last pushed 1mo 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
codspeed
Performance measurement and optimization with CodSpeed: benchmark results, flamegraphs, and regression detection.
reticle
Verify a running web app from the inside. Installs the Reticle skill and registers the Reticle MCP server so the agent can look, act, observe and assert against the real app.
skillgate
Deterministic finish-line gates for AI coding agents — define done once in .skillgate/done.yaml, then block commit, push and publish until it passes. Wires a fail-closed PreToolUse hook, evaluates agent patches in a network-off clone, and keeps CLAUDE.md/AGENTS.md from drifting apart.
agent-done-or-not
Thin Claude Code plugin wrapper that wires proof-of-done Stop hook enforcement.
proofrag
Evaluate a RAG/LLM app: generate a golden set from your docs, run LLM-as-judge + retrieval metrics, and produce a shareable HTML scorecard with a CI gate.
rot-dtd-goal
A /goal engine whose completion is EARNED, not announced. Acceptance criteria are real shell commands; the Stop hook re-runs every one of them and only exits 0 lets a goal finish. Verification defends itself: a sealed integrity ledger re-hashes each criterion before the gate trusts it, a negative-control red team…