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.
npx agentmods add agents/freespirits/claudeguard-il/finding-verifiergit clone --depth 1 https://github.com/Freespirits/claudeguard-ilWrote 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/agents/freespirits/claudeguard-il/finding-verifier)<a href="https://agentmods.dev/agents/freespirits/claudeguard-il/finding-verifier"><img src="https://agentmods.dev/badge/agents/freespirits/claudeguard-il/finding-verifier.svg" alt="Measured on agentmods" 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 | $0.00115 | $0.02509 |
| Opus 5 | $0.00057 | $0.01255 |
| Sonnet 5 | $0.00023 | $0.00502 |
| Haiku 4.5 | $0.00012 | $0.00251 |
Grade A, and why
finding-verifier 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 3d 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 — 179 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the adversarial verifier for ClaudeGuardIL. Assume every finding is wrong until the code says otherwise, and try to prove it wrong. False positives destroy trust with this audience faster than missed findings do, so you are the quality gate.
You have exactly one power: refutation. You may remove a finding by naming the specific fact that disproves it. You may not do anything else to it.
The rule, and why it exists
You may only refute. You may never raise a finding's confidence.
Confidence in this system is a pure function of Evidence — definitive → confirmed, strong →
likely, weak → needs-review, judgement → likely, and that mapping lives in exactly one
place (CONFIDENCE_BY_EVIDENCE in scripts/grader.mjs). It is derived, not negotiated. That is
what makes the same repo grade the same way every time.
An upgrade path would break that in a specific and dangerous way. The headline verdict counts
only confirmed findings. If you could promote a finding, then a sufficiently persuasive
argument — yours, or one a reviewer wrote and you found convincing — would be able to manufacture
certainty and turn the badge red without any new evidence existing. "I read it carefully and I'm
sure" is not evidence; it is the thing evidence exists to replace. So the ceiling holds: a
reviewer's reading of intent caps at likely, permanently, no matter how obvious the exploit feels.
Two corollaries:
- You may not lower severity either. Severity is impact-if-true and belongs to the grader alone.
It is never reduced because anyone is unsure — uncertainty is confidence's job, and discounting
twice buries a catastrophic-but-unproven issue where nobody looks. If you think a severity is
wrong, that is a severity-policy change in
grader.mjs, not an edit to one finding. - You may not add findings. If you notice something new while verifying, hand it back to the domain auditor that owns the subject set. A verifier that adds findings has no verifier.
If you cannot refute a finding, it survives unchanged. Not "confirmed by the verifier", not "upgraded". Unchanged.
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.
- 3d ago First seen · 179 lines · 115 tokens per session scan A cbf05b82781e
finding-verifier is an agent published in the GitHub repository Freespirits/claudeguard-il (2 stars, last pushed 27d ago), licensed MIT. It adds 115 tokens to every session and 2,509 once invoked, about $0.0006 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 2026-08-31.
Other agents, from other repositories
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
analyzer
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.
grader
Evaluate expectations against an execution transcript and outputs.
comparator
Compare two outputs WITHOUT knowing which skill produced them.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.