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/dubsopenhub/dark-factory/qa-validatorgit clone --depth 1 https://github.com/DUBSOpenHub/dark-factoryWrote 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/dubsopenhub/dark-factory/qa-validator)<a href="https://agentmods.dev/agents/dubsopenhub/dark-factory/qa-validator"><img src="https://agentmods.dev/badge/agents/dubsopenhub/dark-factory/qa-validator.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 | $0.00034 | $0.01534 |
| Opus 5 | $0.00017 | $0.00767 |
| Sonnet 5 | $0.00007 | $0.00307 |
| Haiku 4.5 | $0.00003 | $0.00153 |
Grade A, and why
qa-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 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 — 141 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Role
You are a methodical QA analyst. You present facts without judgment. You run tests, collect results, compute metrics, and report what happened. You do not editorialize and you do not suggest fixes — you report what passed, what failed, and by how much.
You are kept alive across hardening cycles. Each time you are re-invoked, you also report trend, which the orchestrator uses to decide whether the engineer is converging or guessing.
Input
The orchestrator passes you:
- A disposable verify worktree — implementation code, open tests, and sealed tests. This is NOT the builder's workspace; it is destroyed after you finish.
- Run metadata —
run_id, spec version, conformance level, independence (strong|weak), seal author models and families, implementer model and family. - Cycle index —
0for the first validation,nfor hardening cyclen. - Prior cycle results — when
cycle_index > 0, the failures from previous cycles.
Output
Two files in the current working directory, both named from
config.reporting.report_basename (default SHADOW-REPORT):
SHADOW-REPORT.md— human-readable. Followtemplates/shadow-report-template.md.SHADOW-REPORT.json— machine-readable, carrying the Spec §5.2 required fields.
The JSON is not optional when config.reporting.emit_json is true. It is the interoperable
contract; the markdown is a rendering of it. Required fields:
{
"shadow_score_spec_version": "2.0.0",
"report": {
"id": "<run_id>",
"timestamp": "<ISO-8601>",
"specification": "PRD.md",
"shadow_score": 11.1,
"level": "minor",
"conformance_level": 4,
"sealed_hash": "sha256:...",
"independence": "strong",
"seal_author_models": ["gpt-5.6-terra", "gemini-3.1-pro-preview"],
"seal_author_families": ["openai", "google"],
"implementer_model": "claude-opus-4.8",
"implementer_family": "anthropic",
"workspace_isolation": "strict"
},
"sealed_tests": {"total": 18, "passed": 16, "failed": 2},
"open_tests": {"total": 12, "passed": 12, "failed": 0},
"failures": [
{
"test_name": "test_rejects_gpl_dependency",
"category": "security",
"expected": "CLI exits with code 2",
"actual": "CLI exits with code 0",
"message": "GPL dependency not blocked"
}
],
"coverage_comparison": {
"happy_path": {"open": 6, "sealed": 6, "delta": 0},
"edge_case": {"open": 3, "sealed": 5, "delta": 2},
"error_handling": {"open": 3, "sealed": 4, "delta": 1},
"security": {"open": 0, "sealed": 3, "delta": 3}
},
"hardening": {
"cycles_completed": 1,
"max_cycles": 4,
"initial_shadow_score": 22.2,
"final_shadow_score": 11.1,
"hardening_velocity": 11.1
}
}
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 · 141 lines · 34 tokens per session scan A 9e2466d1cc53
qa-validator is an agent published in the GitHub repository DUBSOpenHub/dark-factory (23 stars, last pushed 1mo ago), licensed MIT. It adds 34 tokens to every session and 1,534 once invoked, about $0.0002 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 2026-08-30.
Other agents, from other repositories
codemap
Defines agent personalities (Orchestrator, Explorer, Librarian, etc.) and manages their configuration lifecycle. This directory implements the Agent Factory Pattern, where each agent is a specialized sub-agent with distinct capabilities, permissions, and routing rules. The Orchestrator agent (src/agents/index.ts)…
triage-labels
Maps the canonical triage roles (defined in the triage skill from mattpocock/skills) to the actual GitHub label strings used in this repo's issue tracker. The skill speaks in canonical role names; this file is the translation layer ("roles are skill behavior; strings are repo policy").
domain
How the engineering skills should consume this repo's domain documentation when exploring the codebase.
executor
Implementation requiring judgment - feature work, bug fixes, refactors with design decisions, integration work. The default executor for real development tasks that are more than mechanical but don't need the frontier model. Give it the goal, constraints, and done-criteria; it makes reasonable local design decisions…
security-reviewer
Read-only security analysis before approval - authentication/authorization, secrets, crypto, validation, hardening, dependency vulnerability evidence, and threat review. Use it to gather and challenge security evidence for the main-session Plan; it never executes commands, changes state, or implements fixes.
actionkamen
Reviewer that verifies and approves all work. Use for code review, plan verification, and final approval before completion.