Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add Jujitae/invara/plugin install 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/commands/jujitae/invara/assure)<a href="https://agentmods.dev/commands/jujitae/invara/assure"><img src="https://agentmods.dev/badge/commands/jujitae/invara/assure/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/commands/jujitae/invara/assure"><img src="https://agentmods.dev/badge/commands/jujitae/invara/assure.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00050 | $0.01876 |
| Opus 5.5 | $0.00020 | $0.00750 |
| Sonnet 5.5 | $0.00010 | $0.00375 |
| Haiku 4.5 | $0.00005 | $0.00188 |
Grade A, and why
assure scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 145 lines — stays where its author put it; the contents beside it link to each section on GitHub.
The user has a BEFORE and an AFTER — two directories, two commits checked out side by side, a legacy system and its migration — and wants to know whether the AFTER still does what the BEFORE did. You declare how the two are observed; INVARA decides. Quote its words exactly.
0. The tool
export PYTHONPATH="${CLAUDE_PLUGIN_ROOT}/src" # bash / zsh
$env:PYTHONPATH = "${CLAUDE_PLUGIN_ROOT}/src" # PowerShell
Every step is python -m invara assure <step> ... --db .runtime/verify.db.
Runs are local and offline; a request to any host but loopback makes a run
UNVERIFIABLE.
1. Declare the two systems and how they are observed
Write the Equivalence Manifest (.invara/assure.manifest.json). It names
both systems, how an input reaches them, which probes observe them, which
policies relax exact comparison and why, and which claims are mandatory.
{
"schema_version": "invara.assurance.manifest/1",
"session_id": "migration-2026-09-03",
"source_system": {"id": "before", "kind": "process", "command": ["python", "legacy.py"], "root": "$SOURCE_ROOT"},
"target_system": {"id": "after", "kind": "process", "command": ["python", "-m", "newapp"], "root": "$TARGET_ROOT"},
"input_domain": {"kind": "corpus", "delivery": "stdin_json", "corpus": [{"id": "c1", "input": {"...": "..."}}]},
"probes": [
{"id": "cli", "adapter": "process", "capture": ["stderr"], "mandatory": true},
{"id": "out", "adapter": "json", "source": "stdout", "mandatory": true}
],
"policies": [],
"claims": [
{"id": "corpus", "kind": "corpus_equivalence", "mandatory": true},
{"id": "search", "kind": "counterexample_search", "mandatory": true, "params": {"runs": 100, "seconds": 300, "seed": 1}}
]
}
When the whole input space is small and enumerable, declare it and ask for a proof instead of a search:
"input_domain": {"kind": "finite", "delivery": "stdin_json", "finite": {"parameters": {"zone": {"range": [1, 5]}, "express": [false, true]}}},
"claims": [{"id": "domain", "kind": "finite_domain_proof", "mandatory": true}]
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 · 145 lines · 50 tokens per session scan A fe3636e307b8
assure is a command published in the GitHub repository Jujitae/invara (1 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 50 tokens to every session and 1,876 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-09-07.
Other commands, from other repositories
proofrag
Evaluate a RAG/LLM app — generate a golden set, judge it, and produce a scorecard.
bugfix
Quality bug-fix workflow for P1/P2 issues with full TDD cycle.
protocol-consistency
Audit consistency across workstream docs, CLI capabilities, and CI workflows.
test
TDD cycle (Red-Green-Refactor) for test-driven development.
build
Execute a workstream using TDD and guard-enforced workflow.
goal-verify
Run all RoT DTD GOAL acceptance criteria right now and report (does not consume an iteration).