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 fmind/dot --skill ai-red-teamgit clone --depth 1 https://github.com/fmind/dotWrote 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/fmind/dot/ai-red-team)<a href="https://agentmods.dev/skills/fmind/dot/ai-red-team"><img src="https://agentmods.dev/badge/skills/fmind/dot/ai-red-team/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/fmind/dot/ai-red-team"><img src="https://agentmods.dev/badge/skills/fmind/dot/ai-red-team.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 | $0.00019 | $0.00980 |
| Opus 5.5 | $0.00008 | $0.00392 |
| Sonnet 5.5 | $0.00004 | $0.00196 |
| Haiku 4.5 | $0.00002 | $0.00098 |
Grade A, and why
ai-red-team 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 9d 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 — 41 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AI Red Team
Turn a concrete AI attack path into a reproducible assessment and remediation test. threat-model owns architectural analysis, code-security owns source and dependency review, and agent-evaluation owns repeated-trial comparisons. Use PyRIT as the execution framework through a project-local uv environment.
Workflow
- Record the boundary: identify the application revision, model, retrieval sources, actors, tools, and granted authority in the assessment record. Reuse established target, data, action, and cost authorization; proceed autonomously within it. Keep customer evidence in the assessment project.
- Choose plausible scenarios: read attack cases. For each case name the attacker-controlled surface, required access, protected asset, and forbidden outcome. Include legitimate autonomous actions that must continue to succeed.
- Observe the application: capture attempted tool calls and resulting state through the real enforcing boundary. Use disposable records, synthetic secrets, and controlled destinations when these preserve the attack path. A direct model endpoint cannot establish the safety of an application's retrieval or tools.
- Prepare PyRIT: follow PyRIT execution for the released API, target adapter, scenario selection, local smoke, and evidence storage. Freeze package, prompt, dataset, converter, scorer, and application identities before comparative runs.
- Validate the case: prove the intended input reaches the target and that the outcome check detects a deliberately broken control in a fixture. Test a legitimate operation and a rejected operation. Keep target failures, evaluator failures, and actual security failures separate.
- Run the campaign: start with a narrow single-turn case, then add multi-turn or converted cases when the attack path requires them. Declare case count, repeats, concurrency, turn/retry/time/cost limits, cleanup, and stop conditions; retain every attempted trial and its outcome.
- Verify findings: reproduce suspicious results against the application and label confirmed, refuted, and unresolved cases; a model's claim that it accessed a secret is insufficient evidence.
- Remediate and retest: when implementation is authorized, fix the failed runtime control and add a deterministic regression. Repeat the original attack, nearby variants, and legitimate tasks under comparable conditions. Report the remaining uncertainty and the exact tested boundary.
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 9d ago First seen · 41 lines · 19 tokens per session scan A f19974155f72
ai-red-team is a skill published in the GitHub repository fmind/dot (11 stars, last pushed yesterday), licensed MIT. It adds 19 tokens to every session and 980 once invoked, about $0.0001 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 2026-10-08.
Other skills, from other repositories
agent-evaluation
Evaluate stochastic LLM/RAG/model/retrieval/tool agents in trials. Compare baseline/candidate on development/sealed holdouts with calibrated deterministic/model/trace graders; measure reliability, variance, leakage, safety, and cost.
prompt-design
Design production LLM or agent prompt stacks: instructions, tool contracts, examples, outputs, and runtime context. Use for precedence, conflicts, dynamic or untrusted context; prove behavior with agent-evaluation.
dotfiles-bootstrap
Bootstrap a workstation with the dotfiles framework. Takes a GitHub user / owner+repo / explicit clone URL and runs dot init (which shells out to chezmoi) with the right safety prompts. Honors the active agent profile (ask / plan / apply / audit) so it defaults to dry-run in safer modes and full apply in apply.
vibe
Delegate a coding task to a cheap AI model (Mistral Vibe by default, but any provider Vibe knows about — DeepSeek, Gemini Flash, etc.) and supervise the result via git diff. Claude orchestrates, the cheap model codes. Claude consumes 500-1500 tokens per delegation regardless of how many file reads the delegate does…
prompt-engineering
Draft, review, or improve a prompt for an LLM or agent against a research-backed rubric — explicit instructions, instruction/data separation, output contracts, reasoning scaffolds, grounding, verification, and evals. Use when writing a prompt, auditing one that misbehaves, or explaining a prompting principle.
go-stack
Build Go projects, libraries, CLIs, TUIs, web apps, or ADK agents with the standard package layout and pinned tooling.