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 skills add yogsoth-ai/stress-test --skill red-team-truthseekinggit clone --depth 1 https://github.com/yogsoth-ai/stress-testWrote 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/skills/yogsoth-ai/stress-test/red-team-truthseeking)<a href="https://agentmods.dev/skills/yogsoth-ai/stress-test/red-team-truthseeking"><img src="https://agentmods.dev/badge/skills/yogsoth-ai/stress-test/red-team-truthseeking/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/skills/yogsoth-ai/stress-test/red-team-truthseeking"><img src="https://agentmods.dev/badge/skills/yogsoth-ai/stress-test/red-team-truthseeking.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 | $0.00091 | $0.01192 |
| Opus 5 | $0.00046 | $0.00596 |
| Sonnet 5 | $0.00018 | $0.00238 |
| Haiku 4.5 | $0.00009 | $0.00119 |
Grade A, and why
red-team-truthseeking 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 8d 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 — 61 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Red Team (Truth-Seeking Variant)
A retuning of systematic red-teaming. Classic red-teaming enumerates a threat surface, fires attack vectors, and outputs a resilience score (0.0-1.0) plus a list of hardening actions. Two things make that wrong for research: (1) "resilience score" is a defense metric — it rewards un-attackability, the signature of an unfalsifiable claim; (2) "hardening" means patching the artifact to deflect future attacks — exactly the patchwork anti-pattern we reject. This variant keeps the systematic-probing machinery (it is genuinely good at enumeration and coverage) but changes what we enumerate and what we output.
What changed from the original (red-teaming)
| Element | Original (publication/defense) | This variant (truth-seeking) |
|---|---|---|
| Threat surface | Attackable weaknesses | The set of load-bearing CLAIMS (a claim, not a weakness, is the unit) |
| Per-vector goal | Show the artifact can be attacked | Produce the concrete observation/computation that would refute THIS claim |
| Primary output | Resilience score 0.0-1.0 | Refutation-condition per claim (falsifiable? what would break it?) |
| Secondary output | Hardening / mitigation actions | NONE. Findings route to revise/demote/residue, never to patch-to-survive |
| A claim no attack touches | High resilience (good) | UNFALSIFIABLE (RED — worst outcome) |
Core move: assumption → refutation-condition
For each load-bearing claim, the red team does NOT ask "how can I make this look bad?" It asks Platt's strong-inference question: "What is the experiment/observation/computation whose result would force me to abandon this claim?" If a clean such condition exists, the claim is falsifiable and we record it (this is itself the most valuable product — it tells the next round / the sandbox exactly what to measure). If NO such condition can be constructed, the claim is UNFALSIFIABLE and flagged RED.
Execution
1. Threat-surface = load-bearing claim enumeration (threat-surface-mapping, import & repurpose)
Enumerate every claim the artifact LEANS ON — not decorative restatements, the ones that, if false, collapse a downstream conclusion. Sort by load: how many downstream conclusions depend on each. Priority targets are the claims that carry the most weight and the claims stated most confidently relative to their evidence.
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.
- 8d ago First seen · 61 lines · 91 tokens per session scan A 855abd683e22
red-team-truthseeking is a skill published in the GitHub repository yogsoth-ai/stress-test (2 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 91 tokens to every session and 1,192 once invoked, about $0.0005 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 2026-09-03.
Other skills, from other repositories
tournament-judge
Runs a domain-neutral, evidence-based tournament in which independent candidates are reviewed, anonymized, and scored by a blind judge against a rubric fixed before judging. Use when comparing competing plans, designs, implementations, documents, tools, vendors, strategies, or other substantial options; when the user…
sector-rotation
An analysis framework for comparing industries in the Chinese A-share stock market, using business conditions, price momentum, valuation, and money flows. It produces rankings and higher- or lower-allocation suggestions.
browser_cdp
A health check for Python code that produces a score, letter grade, and measures of issues such as complexity, duplication, dead code, dependencies, and architecture.
shogun-screenshot
A screenshot tool for getting images from a computer or web page and then cropping, resizing, or masking sensitive information. Playwright is a browser-automation tool used here to capture web pages.
creating-skills
Guide for creating Claude Code skills following Anthropic's official best practices. Use when user wants to create a new skill, build a skill, write SKILL.md, update an existing skill, or needs skill creation guidelines. Provides structure, frontmatter fields, naming conventions, and new features like dynamic context…
config-validator
Validate AIWG configuration files and project setup for correctness and completeness.