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 CarbeneAI/Forge --skill pyritgit clone --depth 1 https://github.com/CarbeneAI/ForgeWrote 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/carbeneai/forge/pyrit)<a href="https://agentmods.dev/skills/carbeneai/forge/pyrit"><img src="https://agentmods.dev/badge/skills/carbeneai/forge/pyrit/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/carbeneai/forge/pyrit"><img src="https://agentmods.dev/badge/skills/carbeneai/forge/pyrit.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 | $0.00109 | $0.01422 |
| Opus 5 | $0.00055 | $0.00711 |
| Sonnet 5 | $0.00022 | $0.00284 |
| Haiku 4.5 | $0.00011 | $0.00142 |
Grade A, and why
PyRIT 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 6d 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 — 117 lines — stays where its author put it; the contents beside it link to each section on GitHub.
PyRIT — AI Red-Teaming Framework
Microsoft AI Red Team's open-source framework for proactively identifying risks in generative AI systems. Used here for offensive AI engagements (Ehud) and AI governance testing evidence (Daniel CCO — NIST AI RMF MEASURE function, ISO 42001 testing requirements).
Authorized Use Only
PyRIT is for testing AI systems you own or have explicit written permission to test. Engagement scope MUST be defined before running any attack orchestrator. Generated harmful content stays inside the engagement memory store and is purged after report delivery.
Capabilities
- Single-turn attacks — direct prompts, converters (Base64, ROT13, leetspeak, translation, role-play wrappers, etc.)
- Multi-turn attacks — Crescendo, TAP (Tree of Attacks with Pruning), Skeleton Key, custom orchestrators
- Targets — Anthropic Claude, OpenAI, Azure OpenAI, Google, HuggingFace, Ollama (local), custom HTTP/WebSocket endpoints, Playwright-driven web apps
- Scorers — true/false, Likert, classification, Azure AI Content Safety, custom LLM-as-judge
- Datasets — content harms, psychosocial risks, data leakage, prompt injection corpora
- Memory — SQLite (default) for engagement audit trail; Azure SQL optional
- CLI —
pyrit_scan(automated),pyrit_shell(interactive)
Workflow Routing
| Action | Trigger | Behavior |
|---|---|---|
| Quick Scan | "AI red team scan", "test this LLM" | Run pyrit_scan with default scenarios against target |
| Jailbreak Test | "jailbreak test", "test guardrails" | Crescendo or Skeleton Key orchestrator |
| Prompt Injection | "prompt injection test", "test for injection" | Single-turn injection corpus + scorer |
| Multi-turn TAP | "tree of attacks", "TAP attack", "deep red team" | TAP orchestrator with adversarial LLM |
| Custom Engagement | "custom AI red team", "build PyRIT scenario" | Drop into Python with tools/run-pyrit.sh |
| Governance Evidence | "AI risk assessment", "NIST AI RMF MEASURE", "ISO 42001 testing" | Run scoped scan, output to Obsidian for audit binder |
What ships with it
1 file 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.
- 6d ago First seen · 117 lines · 109 tokens per session scan A 71e91f6e08d2
PyRIT is a skill published in the GitHub repository CarbeneAI/Forge (9 stars, last pushed 1mo ago), licensed MIT. It adds 109 tokens to every session and 1,422 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
ai-ethics-governance
../../../platform-ai/ai-ethics-governance/SKILL.md.
ai-red-teaming
../../../red-team/ai-red-teaming/SKILL.md.
update-llm-models
Autonomously research and update the default lightweight LLM models in src/core/llm.ts using web search to ensure the most current versions are used.
vector-and-embedding-weaknesses
Hunt vector / embedding weaknesses (OWASP LLM08:2025) — adversarial inputs against the RAG / similarity layer that cause cross-tenant leak, embedding-inversion privacy loss, semantic confusion, and retriever-driven prompt injection.
llm-redteam-overview
LLM red team category — full AATMF v3 tactic coverage (T01–T15). Routing skill: read this first to identify which tactic applies, then load the matching sub-skill. Maps to MITRE ATLAS where overlap exists.
aatmf-t06-training-poisoning
AATMF T6 — Training & Feedback Poisoning. Data poisoning, RLHF reward hacks, fine-tune-time exfil, embedding poisoning.