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 forefy/.context --skill llm-fuzz-toolsgit clone --depth 1 https://github.com/forefy/.contextWrote 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/forefy/.context/llm-fuzz-tools)<a href="https://agentmods.dev/skills/forefy/.context/llm-fuzz-tools"><img src="https://agentmods.dev/badge/skills/forefy/.context/llm-fuzz-tools/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/forefy/.context/llm-fuzz-tools"><img src="https://agentmods.dev/badge/skills/forefy/.context/llm-fuzz-tools.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 | $0.00050 | $0.01499 |
| Opus 5 | $0.00025 | $0.00749 |
| Sonnet 5 | $0.00010 | $0.00300 |
| Haiku 4.5 | $0.00005 | $0.00150 |
Grade A, and why
llm-fuzz-tools 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 4d 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 — 123 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Contents
- Scope, authorization & cost
- Phase 1 - pick two tools, not five
- Phase 2 - run them as background jobs
- Phase 3 - normalize
- Phase 4 - dedupe and triage
- False-positive gates
- Output
- Reference files:
references/tool-coverage.md,references/finding-schema.md
Scope, authorization & cost
Only scan an endpoint you own or are contractually engaged to test.
Cost is a first-class constraint here, not a footnote. Every other skill in this set spends your time; this one spends the client's money. A single broad scanner is on the order of 200 probes, each a paid completion, multiplied by every provider binding you point it at, multiplied by any mutation layer. Two tools with overlapping corpora double the bill for a fraction of the coverage.
Before the first run:
- Get a stated ceiling in writing, in currency, not in probe count.
- Confirm whether the key you were given is production or a test project. Scanner traffic on a production key distorts the client's own usage metrics and can trip their abuse detection.
- Check the endpoint's rate limits. A run that dies at 40 percent looks like a clean result.
- Run one tool's smallest probe subset first, measure actual spend per probe, and extrapolate before committing to the full matrix.
Blast radius: this sends adversarial prompts to a metered third-party API and stores the responses. Some responses will be harmful content by design. Treat the artifact store as sensitive.
Phase 1 - pick two tools, not five
The instinct is to run everything and sum the results. That produces a bigger number, not better coverage: the broad corpus scanners carry substantially the same public jailbreak and injection sets, so the second one mostly re-finds the first one's hits at full price.
references/tool-coverage.md splits the field by what each tool uniquely does. The short version:
- one broad single-shot corpus scanner (they overlap heavily with each other; pick on runtime and provider bindings, not on probe count)
- plus a multi-turn attack framework, if adversarial conversation is in scope - single-shot corpora structurally cannot find what only emerges over several turns
- plus an eval harness, only if you own the application and want the result wired into CI
What ships with it
2 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.
- 4d ago First seen · 123 lines · 50 tokens per session scan A d8735ad4bf7c
llm-fuzz-tools is a skill published in the GitHub repository forefy/.context (146 stars, last pushed 3d ago), licensed MIT. It adds 50 tokens to every session and 1,499 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 2026-09-08.
Other skills, from other repositories
LQF_Machine_Learning_Expert_Guide
LQF Machine Learning Expert Guide - Routed skill for ML/Statistical Modeling with Critical Discussion Mode. Triggers on: machine learning, modeling, prediction, training, classification, regression, clustering, deep learning, neural network, model evaluation, feature engineering, hyperparameter tuning, overfitting…
imaging-data-commons
Query and download public cancer imaging data from NCI Imaging Data Commons using idc-index. Use for accessing large-scale radiology (CT, MR, PET) and pathology datasets for AI training or research. No authentication required. Query by metadata, visualize in browser, check licenses.
similarity-search-patterns
Implement efficient similarity search with vector databases. Use when building semantic search, implementing nearest neighbor queries, or optimizing retrieval performance.
embedding-strategies
Select and optimize embedding models for semantic search and RAG applications. Use when choosing embedding models, implementing chunking strategies, or optimizing embedding quality for specific domains.
nowait-reasoning-optimizer
Implements the NOWAIT technique for efficient reasoning in R1-style LLMs. Use when optimizing inference of reasoning models (QwQ, DeepSeek-R1, Phi4-Reasoning, Qwen3, Kimi-VL, QvQ), reducing chain-of-thought token usage by 27-51% while preserving accuracy. Triggers on "optimize reasoning", "reduce thinking tokens"…
explaining-machine-learning-models
Explain trained machine learning models through feature attribution, local explanations, and behavior summaries. Use as an explicit/manual helper once a model already exists, not for training ownership, leakage auditing, or general ML strategy selection.