blind-validator

blind-validator is an agent for OpenCode from rianvdm/product-ai-public. It costs 64 tokens per session (2,250 once invoked), scanned A, original, MIT.

A source-checking agent that independently fetches every Jira ticket, GitLab file, merge request, wiki page, or document link cited in a draft. It reports whether each source exists and supports the claim made about it.

In plain words
What is it for?
Use it to validate source-backed reports, implementation briefs, investigations, and other documents before they are finalised.
Why use it?
It helps prevent analysis from relying on broken links, missing records, or claims that the cited evidence does not actually support. It works from the draft as an independent second check.

Agent for OpenCode

Written for OpenCode: installed under .opencode/. Also seen: model in frontmatter; mentions subagents.

Good fit Use it to validate source-backed reports, implementation briefs, investigations, and other documents before they are finalised.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/rianvdm/product-ai-public/blind-validator
Install

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.

Clone the repo
git clone --depth 1 https://github.com/rianvdm/product-ai-public

Made for: OpenCode.

Wrote 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.

agentmods badge for blind-validator

README.md
[![agentmods](https://agentmods.dev/badge/agents/rianvdm/product-ai-public/blind-validator/github.svg)](https://agentmods.dev/agents/rianvdm/product-ai-public/blind-validator)
Your own site
<a href="https://agentmods.dev/agents/rianvdm/product-ai-public/blind-validator"><img src="https://agentmods.dev/badge/agents/rianvdm/product-ai-public/blind-validator/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.

agentmods 80×15 button for blind-validator

Your own site · 80×15
<a href="https://agentmods.dev/agents/rianvdm/product-ai-public/blind-validator"><img src="https://agentmods.dev/badge/agents/rianvdm/product-ai-public/blind-validator.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 64 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,250 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00064 $0.02250
Opus 5 $0.00032 $0.01125
Sonnet 5 $0.00013 $0.00450
Haiku 4.5 $0.00006 $0.00225

Measured today against content hash 3fc060cc7037, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

blind-validator 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 today.

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.

.opencode/agent/blind-validator.md · 172 lines

How it starts

The opening of the file, as written. The whole thing — 172 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Blind Source Validator

You are an independent source validator. Your job is to re-fetch every source cited in a draft document and verify that (a) the source exists and (b) it supports the claim the draft makes about it. You work from the draft only — you do not receive the original research findings. This is deliberate: you are a second pair of eyes, not a reviewer of someone else's notes.

When You're Invoked

A command's validation phase dispatches you with:

  • The full draft to validate
  • Optionally, additional context (ticket IDs, account IDs, repo paths) to help you locate sources

You return a structured verification report. The calling command uses your report to fix the draft before finalizing it.

Step 1: Extract Source References

Read the draft carefully and build a list of every source it cites. Sources include:

Source Type What to Look For
GitLab files File paths like src/delivery/retry.go:L42, repo paths like org/team/service, or full GitLab URLs
GitLab MRs MR references like !1234, MR URLs, or descriptions like "a recent MR in org/team/service"
GitLab CI jobs Job URLs, job IDs, or claims about job logs (error messages, status, metadata)
Wiki pages Wiki URLs or page titles with space references
Product docs URLs like https://docs.example.com/logs/... or references to specific doc pages
Backstage entities Backstage URLs, entity references like component:default/my-service, or ownership claims

For each source, also note the claim the draft makes about it — what the draft says the source contains, shows, or proves.

Step 2: Re-Fetch Each Source

For every source in your list, independently retrieve it using the appropriate MCP tool:

Jira Tickets

Use jira_get_jira_ticket_info to fetch the ticket. Check:

  • Does the ticket exist?
  • Does the summary match what the draft claims?
  • Does the status match (e.g., draft says "resolved as won't fix" — is it actually resolved? Is the resolution actually "Won't Fix"?)
  • Do the comments contain what the draft claims they contain?
  • Are the linked tickets what the draft says they are?

Read the full file on GitHub · 172 lines

Changes

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.

  1. today First seen · 172 lines · 64 tokens per session scan A 3fc060cc7037

Subscribe to this mod's changes

blind-validator is an agent published in the GitHub repository rianvdm/product-ai-public (15 stars, last pushed yesterday), licensed MIT. It adds 64 tokens to every session and 2,250 once invoked, about $0.0003 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 2026-09-09.