elementalsouls/Claude-AgentHunter

記録された最新のソース活動: 2026年10月19日 19:35

SkillsMP カタログ更新: 2026年10月29日 18:30

収集済み skills: 24

GitHub スター: 0
GitHub フォーク: 0

AI アシスタントでインストール

このプロンプトをコピーして、使用中の AI アシスタントに送ってください。

Follow https://skillsmp.com/skill-install/prompt.md to install Agent Skills from https://github.com/elementalsouls/Claude-AgentHunter.

リポジトリの README

README の元ファイル：README.md

README の元ファイルを読む

README 全文を表示

![Claude-AgentHunter banner](assets/banner-v2.svg) # Claude-AgentHunter > **Pentest AI agents — and prove both the findings and the absences.** A skill bundle plus a runnable engine for red-teaming LLM agents: chatbots, RAG assistants, tool-calling copilots, MCP servers, and autonomous ReAct agents · **24 skills** (23 hunt + methodology) · 4 detection arms · 24/24 lab gates · 7 target environments · FN 0 over 949 field runs · loads into Claude Code, OpenCode, Hermes, Codex CLI and AntiGravity. Built by **[Sachin Sharma](https://www.linkedin.com/in/sachinsharma8080/)** — Bug Hunting & GenAI Security Research. --- Sibling to [Claude-BugHunter](https://github.com/elementalsouls/Claude-BugHunter) — same shape, different quarry. BugHunter hunts web apps; Claude-AgentHunter hunts the agents themselves. ``` 24 skills · 4 detection arms · 24/24 lab gates · 7 target environments · FN 0 over 949 runs ``` ## What is this A drop-in skill bundle for the [Claude Code skills system](https://docs.claude.com/en/docs/claude-code/skills) (and OpenCode, Hermes, Codex CLI, AntiGravity). Install once and the harness stops guessing at agent security and starts running a disciplined pentest: it fingerprints the target agent, picks the right `hunt-agent-*` skills by surface, runs their stdlib-only probes, scores the traces with a 4-arm engine, and gates every finding so a fluent answer never ships as a leak and a silent probe never ships as "clean". Describe the target in plain English and the relevant skills auto-load — no invocation by name. The whole flow runs as one command, `/hunt-agent`, or stage by stage (see [Commands](#commands)). ## Why it exists An agent generates text; it does not retrieve state. That creates two failure modes a web pentest never has to worry about, and **Claude-AgentHunter is built around catching both**: - A fluent answer looks identical to a real leak → the **False-Positive gate**. - A silent probe looks identical to a secure target → the **Negative-Result gate**. The second one is the hard one. Across this project's own development, **ten probe defects each reported a compromise as clean** — a similarity gate that scored a real leak 0.03, an HTTP 500 counted as a refusal that buried a confirmed bypass, a token floor that hid the forgery that leaked the flag. Every one looked like a passing test. Both gates live in `engine/gates.py` and ship in every skill. ## How the skills work Each skill is `skills/<name>/SKILL.md` — Agent-Skills markdown with frontmatter (`name`, `description`, `sources`, `validated_against`, `report_count` — the count of in-house field-validated reports behind the skill, 0 for reference-only). Harnesses topic-match on the description; describing your test in plain English loads the relevant skills, no explicit invocation. Unlike static methodology skills, each one carries a **Runnable Probe** that executes against a live agent and a **`# LETHAL WHEN:`** line tying its output to the gate. Catalog: [`docs/skills.md`](docs/skills.md). ## The engine (what connects and merges) `engine/` is the deterministic core. It normalises any target into one trace shape (`engine/adapters/`), runs four complementary detection arms over it, and gates the result: | arm | detects | blind to | |---|---|---| | **A · taint** | an attacker-planted value reaching a tool argument | injections that reuse data the agent already holds | | **B · unrequested action** | a state-changing call the clean baselines never made | wording drift (needs baselines scaled to argument entropy) | | **C · output echo** | injected content surfacing in the agent's own voice | paraphrase; quoted data (stripped) | | **D · provenance** | the agent acting on a value only untrusted content supplied | nothing structural — the general case | No arm is sufficient alone. Taint is 64/64 on one suite and **0/39** on another. A target is clean only when all arms agree **and their controls fired**. ```sh python3 engine/score.py agentdojo <runs_dir> <suite> --baselines d1,d2 ``` ## Commands The pipeline runs end to end as `/hunt-agent`, or one stage at a time: | command | stage | |---|---| | `/characterize-agent <target>` | fingerprint visibility · allowlist · egress · oracle · auth — before any payload | | `/hunt-agent <target>` | full workflow: characterize → select skills → probe → score → gate | | `/score-agent <runs_dir> <suite>` | run the 4 detection arms over captured traces + apply the gates | | `/validate-agent` | gate one finding through the FP + Negative-Result gates | | `/report-agent` | write the submission-ready finding, scored by the sink it reached | ## Install ```sh git clone https://github.com/elementalsouls/Claude-AgentHunter && cd Claude-AgentHunter ./scripts/install.sh --all # detects Claude Code, OpenCode, Hermes, Codex, AntiGravity ./scripts/install.sh --claude # or one harness ``` Or as a Claude Code plugin — namespaced, no copy, updates on version bump: ```sh /plugin marketplace add elementalsouls/Claude-AgentHunter /plugin install claude-agenthunter@elementalsouls ``` ## Runs on multiple agent harnesses ![harness routing](assets/harness-routing.svg) The skills are plain [Agent Skills](https://docs.claude.com/en/docs/claude-code/skills) — the same `SKILL.md` format that Claude Code · OpenCode · OpenAI Codex CLI · Hermes Agent · Google AntiGravity all load. One command installs them everywhere: ```sh # macOS / Linux ./scripts/install.sh --all # Windows (PowerShell) pwsh ./scripts/install.ps1 -All ``` | harness | target skill directory | flag | |---|---|---| | **Claude Code** (baseline) | `~/.claude/skills/` | *(default)* | | **OpenCode** | reads `~/.claude/skills/` & `~/.agents/skills/` | *(default)* \| `--agents` | | **OpenAI Codex CLI** | `~/.agents/skills/` | `--agents` | | **Hermes Agent** | `~/.hermes/skills/` | `--hermes` | | **Google AntiGravity** | `~/.gemini/config/skills/` | `--antigravity` | `--all` detects the installed harnesses and copies the skills to each path. The full **knowledge** layer (24 skills) ports to all five; the slash commands and the `engine/` + `lab/` tooling stay Claude-Code / standalone by design. ## Proof it works — `lab/` and `field/` - **`lab/`** — 24 deterministic exploit gates (`python3 lab/run_lab.py`), no model, no API key. Each encodes a vulnerability class and its control; `[CONFIRMED]` means re-runnable, not "the model said so". - **`field/`** — evidence from live/benchmark targets: AgentDojo (4 suites, 949 runs, FN 0), DVMCP (10/10), DVLA (ReAct forgery → SQL chain), InjecAgent (arm D 7/7 indirect), a real MCP target, ARKX. ## Documentation - [`USAGE.md`](USAGE.md) — the workflow, the engine, the lab, field evidence, end to end. - [`INSTALL.md`](INSTALL.md) — plugin + copy install across five harnesses. - [`docs/skills.md`](docs/skills.md) — the 23 hunt-agent skills catalog (+ the `ai-agent-methodology` orchestrator = 24 total). - [`docs/THREAT-MODEL.md`](docs/THREAT-MODEL.md) · [`docs/COVERAGE-MATRIX.md`](docs/COVERAGE-MATRIX.md) · [`docs/RESEARCH-2026.md`](docs/RESEARCH-2026.md) · [`docs/DEFENSES-AND-DETECTION.md`](docs/DEFENSES-AND-DETECTION.md) · [`docs/REFERENCES.md`](docs/REFERENCES.md) - [`CONTRIBUTING.md`](CONTRIBUTING.md) · [`CHANGELOG.md`](CHANGELOG.md) ## Authorized use only For testing agents you own or are explicitly authorized to assess. The probes send live payloads and the lab targets are intentionally vulnerable. See [`SECURITY.md`](SECURITY.md).

リポジトリの README を読み込んでいます…

このリポジトリの skills

分類待ち

未分類

Temporal-constraint bypass in AI-agent tools — future-dating, weekend/maintenance-window evasion, rate-reset manipulation. Time gates enforced as prompt instructions are not real time locks. A TOCTOU variant: cross-check hunt-agent-race-condition before…

原文の言語: 英語

2026-10-19

収集済み skill 24 件中 24 件を表示しています。