[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7ASGDNBRTX1BA5FHMP7B36H":3,"skill-same:01M7ASGDNBRTX1BA5FHMP7B36H":48},{"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":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":29,"ownerHtmlUrl":30,"ownerBlogUrl":31,"ownerDescription":32,"updatedAt":33,"install":34,"resources":47},"01M7ASGDNBRTX1BA5FHMP7B36H","tor-protocol-circuits-and-cell-cryptography",null,"Use when reasoning about how Tor actually works on the wire — cells, channels, the telescoping circuit handshake, ntor and ntor-v3, the tor1 → CGO cell-crypto migration, SENDME flow control, congestion control and Conflux — or when correcting a claim about what Tor promises. Covers §1–§5. Companion to the other Tor and onion-network skills.","the-vibey-project\u002Fvibey","the-vibey-project","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F329298758?v=4","01M7AJHGZ9722CTJ4JNQZVRH0Q","the vibey project","01M7AJHGZVMG2K7TGXQ6XFHVG3",5,"MIT","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Ftor-and-onion-networks\u002Fskills\u002Ftor-protocol-circuits-and-cell-cryptography\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: tor-protocol-circuits-and-cell-cryptography\ndescription: \"Use when reasoning about how Tor actually works on the wire — cells, channels, the telescoping circuit handshake, ntor and ntor-v3, the tor1 → CGO cell-crypto migration, SENDME flow control, congestion control and Conflux — or when correcting a claim about what Tor promises. Covers §1–§5. Companion to the other Tor and onion-network skills.\"\n---\n\n# Tor: What It Promises, How Circuits Are Built, and What Encrypts a Cell\n\n> **Part 1 of 8** of the *Tor and Onion Networks* reference (plugin `tor-and-onion-networks`),\n> covering §1–§5. Sibling skills: `tor-network-consensus-guards-and-paths` (§6–§7),\n> `tor-onion-services-and-hardening` (§8–§9),\n> `tor-censorship-circumvention-and-bridges` (§10),\n> `tor-attack-literature-and-threat-model` (§11),\n> `tor-ecosystem-alternatives-and-governance` (§12, §14–§16),\n> `tor-building-on-tor-software-patterns` (§13.1–§13.4, handbook §1),\n> `tor-running-relays-bridges-and-hardware` (§13.5, handbook §2–§6).\n> Section numbers are shared across the set; a reference written as §N → `skill` points into\n> that sibling skill.\n>\n> **Currency:** compiled 16 September 2026. Protocol fundamentals (cell formats, handshake\n> math) are stable and change slowly. **Version numbers and deployment status are dated** —\n> the tor1 → CGO transition in §4 and the Conflux rollout in §5 are mid-flight, so verify\n> against \u003Chttps:\u002F\u002Fspec.torproject.org\u002F> before relying on them months from now.\n\n> **⚠️ A low-latency anonymity overlay: three independently operated relays and layered\n> encryption, so no single party links who you are to what you are doing.**\n>\n> **⚠️ GOTCHA** boxes mark where the mental model people carry is wrong in ways that cause\n> real deanonymization or broken tooling.\n>\n> **The three ideas that organize this document:**\n> 1. **⚠️ TOR SEPARATES IDENTIFICATION FROM ROUTING** (§1, §3). **The guard knows who you\n>    are but not what you do; the exit knows what you do but not who you are. Every design\n>    decision downstream — guards, families, path restrictions — exists to keep those two\n>    facts in different hands.**\n> 2. **⚠️ THE TLS BETWEEN RELAYS IS NOT THE ANONYMITY LAYER** (§3.3). **Hop-to-hop TLS\n>    protects against someone sitting between relays. The anonymity comes from the\n>    onion-encrypted relay cells riding inside it. Confusing the two is the most common\n>    architectural misreading of Tor.**\n> 3. **⚠️ TOR'S PERFORMANCE PROBLEM WAS NEVER CRYPTOGRAPHIC — IT WAS QUEUEING** (§5).\n>    **Twenty years of latency complaints came from unmanaged queues at slow relays, which\n>    is why the 2022–2026 work is congestion control and traffic splitting, not faster\n>    ciphers.**\n\n---\n\n## §1. What Tor is, what it promises, and what it does not\n\nTor is a low-latency anonymity overlay network: roughly 9,200 volunteer relays and 2,300\nbridges carrying traffic for several million daily users (network-size and user CSV data\nfrom [Tor Metrics](https:\u002F\u002Fmetrics.torproject.org\u002F), 8–14 Sept 2026; see §6 →\n`tor-network-consensus-guards-and-paths` for detail). A client's traffic is routed through a\nchain (\"circuit\") of three independently operated relays, with layered (\"onion\") encryption\nsuch that:\n\n- the **first relay (guard)** knows *who you are* but not what you're doing;\n- the **middle relay** knows neither;\n- the **exit relay** (or onion-service rendezvous point) knows *what is being done* but not\n  *by whom*.\n\nThe design goal is **unlinkability**: no single entity — relay, ISP, destination server —\nshould be able to connect a user's identity to their activity. Tor provides *transport-layer*\nanonymity: it hides network endpoints, not the content or the behavior inside the stream.\n\n### ⚠️ What Tor deliberately does not promise\n\nPer its own threat model:\n\n- **Protection against a global passive adversary.** Anyone who can watch both ends of a\n  circuit (e.g., your ISP *and* the destination's network) can confirm relationships by\n  timing\u002Fvolume correlation. Tor is low-latency by design; defeating end-to-end correlation\n  is what high-latency mix networks try to do instead (§14 →\n  `tor-ecosystem-alternatives-and-governance`).\n- **Confidentiality past the exit relay.** Exit traffic to a clearnet site leaves the Tor\n  envelope. Use end-to-end encryption (HTTPS) as usual; Tor Browser bundles HTTPS-only mode\n  partly for this reason.\n- **Application-level anonymity.** Logging into your personal account over Tor, or running a\n  browser that leaks identifying fingerprints, defeats the network layer. This is why the Tor\n  Project ships a hardened browser rather than telling people to point any app at a SOCKS\n  proxy.\n- **Invisibility.** A censor can see that you *are using* Tor unless you use a pluggable\n  transport (§10 → `tor-censorship-circumvention-and-bridges`). Tor hides destinations, not\n  the fact of its use.\n\nA useful mental model: Tor gives you **three organizational trust boundaries** per circuit\nplus layered crypto, and then throws engineering effort (guards, congestion control, padding\nresearch, anti-enumeration transports) at shrinking the residual deanonymization surface.\n\n---\n\n## §2. History and lineage\n\n| Era | Development |\n|---|---|\n| Mid-1990s | Onion routing invented at the U.S. Naval Research Laboratory (Paul Syverson, David Goldschlag, Michael Reed). The insight: separate *identification* from *routing* by nesting encrypted hops. |\n| 2002 | First-generation onion routing prototype deployed at NRL. |\n| 2003–2004 | Roger Dingledine, Nick Mathewson, and Paul Syverson build the second-generation system — \"Tor\" — fixing the first gen's problems (no directory system, replay\u002Fintegrity weaknesses, usability). The USENIX Security 2004 paper *Tor: The Second-Generation Onion Router* remains the canonical design document. |\n| 2004–2006 | EFF funds early development; the network opens to the public. |\n| Dec 2006 | The Tor Project, Inc. founded as a U.S. 501(c)(3) nonprofit. |\n| 2008–2011 | Tor Browser Bundle (originally \"Torbutton\") makes the system usable by non-experts; bridges introduced to counter blocking in Iran and China. |\n| 2013–2014 | Snowden-era scrutiny and usage spike; ntor replaces the old RSA\u002FTAP handshakes for circuit extension; hidden-service deanonymization research (and at least one real-world attack campaign that surfaced in 2014) catalyzes the v3 redesign. |\n| 2015–2018 | `.onion` registered as a special-use domain (RFC 7686); v3 onion service spec written (224\u002F225 proposals era); ed25519 relay identities roll out. |\n| 2017 | Directory authority diversity pushed; Tor Browser 7 based on Firefox ESR 52 brings multiprocess. |\n| Oct 2021 | v2 onion services (16-char, RSA-1024, SHA-1-derived) **removed from the network**. All onion services are now v3 (56-char). |\n| 2022–2023 | Congestion control (proposal 324) ships in C tor 0.4.7 and is tuned through 0.4.8; Conflux traffic-splitting (proposal 329) developed. |\n| 2023 | Onion-service proof-of-work DoS defense (proposal 327) ships in tor 0.4.8 (Aug 2023). |\n| 2024 | Conflux enabled network-wide in C tor (controlled by consensus parameter); Arti (the Rust rewrite) reaches **client-side feature parity** and gets native vanguards. Tor Project and Tails merge organizations. WebTunnel transport introduced. |\n| 2025 | Tor 0.4.9 alphas introduce **CGO cell crypto** and **happy families**; Tor Browser 15 (Firefox ESR 140); Arti 2.x series begins; Tor VPN beta (Arti-based Android VPN); Conflux hardening continues; Tails 7 (Debian 13) and Qubes 4.3 ship. |\n| 2026 (as of Sept) | C tor 0.4.9.5 is the first stable 0.4.9 (12 Feb 2026); the 0.4.8 series is on the deprecation path. Arti 2.6.0 (1 Sept 2026) makes CGO and congestion control always-on, and pushes toward **Arti-as-relay** (directory mirror, DNS streams, channel auth). |\n\n**⚠️ The pattern to internalize:** **C tor is in managed decline, Arti is the future.** In\n2026 the frontier work is relay\u002Fdirectory-authority support in Arti + post-SHA-1 cell crypto\n(CGO) + congestion\u002Fflow-control modernization.\n\n---\n\n## §3. The core protocol: cells, circuits, and the telescoping handshake\n\nEverything in Tor happens in **cells** — fixed-size messages multiplexed over TLS connections\n(\"channels\") between relays. The classic cell is 514 bytes on the wire: a 2-byte circuit ID\n(4 bytes on modern link protocols), a 1-byte command, and a 509-byte payload. Variable-length\ncells exist for protocol negotiation and directory data.\n\nCell commands split into two families:\n\n- **Channel-level:** `CREATE`\u002F`CREATED` (now `CREATE2`\u002F`CREATED2`), `DESTROY`, `NETINFO`,\n  `CERTS`, `AUTH_CHALLENGE`\u002F`AUTHENTICATE`, `PADDING` — negotiated directly with the\n  immediate neighbor.\n- **Circuit-level:** `RELAY` (and `RELAY_EARLY` near circuit creation) — onion-encrypted\n  payloads that tunnel *through* the channel, understood only by whichever hop peels that\n  layer.\n\n### §3.1 How a circuit is built (\"telescoping\")\n\n1. **Client → Guard.** The client performs an authenticated ntor key exchange (`CREATE2`)\n   with its chosen guard. Both sides now share a symmetric key pair (forward\u002Fbackward\n   digests) for that hop. Result: circuit hop 1.\n2. **Extend to hop 2.** The client builds an `EXTEND2` relay cell containing the link\n   specifiers of the middle relay plus a fresh ntor handshake blob, onion-encrypted for the\n   guard. The guard unwraps, opens (or reuses) a channel to the middle, forwards the\n   handshake, and returns the middle's `CREATED2` response inside an `EXTENDED2` relay cell.\n   The client now has shared keys with hop 2 as well — negotiated *through* hop 1 without hop\n   1 learning the hop-2 key.\n3. **Extend to hop 3 (exit).** Same procedure, now layered through hops 1 and 2.\n\nEach hop strips one layer of symmetric crypto from outbound cells and adds one layer to\ninbound cells — the \"onion.\" The exit decrypts the final layer and forwards plaintext to the\ndestination (for client circuits) or the service.\n\n```\n\u002F\u002F Outbound (client → destination)\nClient builds:  Layer1(Layer2(Layer3(payload)))\nGuard:  strips Layer1 → forwards Layer2(Layer3(payload)) to middle\nMiddle: strips Layer2 → forwards Layer3(payload)         to exit\nExit:   strips Layer3 → forwards payload                 to destination\n\n\u002F\u002F Inbound (destination → client) reverses: exit adds Layer3, middle adds Layer2,\n\u002F\u002F guard adds Layer1, and the client — holding keys for all three hops — strips all three.\n```\n\nThis is the whole reason the client must hold keys for every hop and each hop holds keys only\nfor itself: the client is the only party that can construct or unwrap the full stack.\n\nTwo safety rails worth knowing:\n\n- **`RELAY_EARLY` cells** bound circuit construction: only the first N (8) relay cells in a\n  circuit's life may extend it, which prevents infinitely long circuits being used as a\n  resource-exhaustion vector and bounds certain tagging\u002Fconfirmation games.\n- Circuit IDs scope cells within a channel; `NETINFO` cells exchange timing\u002Faddress data used\n  for NAT traversal and address confirmation.\n\n### §3.2 Streams\n\nOnce a circuit exists, the client attaches **streams** (TCP connections) to it with\n`RELAY BEGIN`; the exit answers `CONNECTED`. Data flows as `RELAY DATA` cells; `RELAY END`\ncloses a stream. Tor multiplexes many streams over one circuit, and one client's circuits may\ncarry arbitrarily many streams — but see **stream isolation** (§13.1 →\n`tor-building-on-tor-software-patterns`) for why Tor Browser splits streams by first-party domain\nanyway.\n\n### §3.3 ⚠️ GOTCHA: where the TLS ends\n\nA subtlety newcomers miss: the hop-to-hop transport is TLS over TCP, but that TLS only\nprotects against attackers sitting *between* relays — it is **not** the anonymity layer. The\nanonymity comes from the onion-encrypted relay cells riding inside. Relay channels\nauthenticate with ed25519 identities (since the 0.3.0 era), and as of C tor 0.4.9 (Feb 2026)\nthe last pre-ed25519 link authentication methods (v1\u002Fv2 handshakes, the old\n`RSA-SHA256-TLSSecret` auth) were **removed**, along with fake-TLS cipher advertisement code.\n\n---\n\n## §4. Circuit cryptography: from TAP to ntor to CGO\n\n### §4.1 The handshake stack\n\n| Generation | Used for | Status (Sept 2026) |\n|---|---|---|\n| TAP | RSA-1024 + DH-1024 first-hop create, old extend | **Fully removed** from C tor in 0.4.9 (2026) |\n| ntor | Curve25519\u002FX25519 one-way-authenticated key exchange for `CREATE2`\u002F`EXTEND2`; variant `hs-ntor` for onion-service intro\u002Frendezvous handshakes | Standard, universal |\n| **ntor-v3** | Extended ntor with authenticated client→server data in the handshake (needed by CGO's crypto negotiation) | Rolling out; required for CGO negotiation |\n\nThe ntor design (Goldberg–Stebila–Ustaoglu, 2012–2013) gives **one-way authentication in a\nsingle round trip**: the server proves identity via its long-term key, the client stays\nanonymous, and both derive forward-secret session keys. The cost that motivated it — the old\nTAP handshake needed full RSA operations per hop and was CPU-dominant for relays — is why\nmodern Tor scales at all.\n\n### §4.2 The cell crypto itself: \"tor1\" and its replacement\n\nFor two decades the relay-level symmetric crypto (\"tor1\") was AES-128 in counter mode plus a\n4-byte running SHA-1 digest for integrity. Cryptographers flagged its problems for years:\n\n- **Tagging attacks.** CTR mode is malleable: a hop (or observer who controls a hop) can flip\n  predictable bits in a cell and watch the effect emerge at another hop — a marked cell\n  becomes a tracking beacon through the circuit.\n- **Weak integrity.** Four bytes of digest is ~32-bit authentication, and the underlying hash\n  was SHA-1.\n- **No forward secrecy at the cell layer** — keys stayed constant for the life of the circuit.\n\nThe replacement, **Counter Galois Onion (CGO)**, designed by Degabriele, Melloni, Münch and\nStam (proposal 308, 2019; revised and instantiated as **proposal 359, \"CGO Redux\"**), is the\nbiggest change to Tor's wire format since v3 onion services:\n\n- Built on **UIV+**, a *rugged pseudorandom permutation*: encryption is non-malleable (any\n  tamper, and that cell *plus every subsequent cell in that direction* becomes unrecoverable\n  — killing tagging), while relays process cells via a deliberately malleable decrypt path.\n- Primitives are AES-128 (counter PRF) + **POLYVAL** universal hashing — chosen to exploit\n  AES-NI and PCLMULQDQ hardware instructions, so the security upgrade is also fast on modern\n  server CPUs.\n- Cell format becomes a **16-byte authenticator tag + 493-byte body**; the running SHA-1\n  digest and the `Recognized` field disappear (SENDMEs now bind to the cell tag).\n- Keys **evolve per cell** (`UPDATE_UIV`), so old cells can't be re-decrypted later even with\n  the current key — forward secrecy at the cell layer.\n- **Deployment status (Sept 2026):** implemented in C tor 0.4.9 (clients and relays negotiate\n  it) and in Arti, where CGO + congestion control became **always-on in Arti 2.6.0 (1 Sept\n  2026)**, after being marked stable in Arti 2.5.0 (30 June 2026). It requires ntor-v3 and the\n  new FlowCtrl congestion signaling — which is why it arrived together with the flow-control\n  redesign below. Onion-service-circuit CGO was in experimental stabilization in Arti 2.5.x\n  (mid-2026).\n\n> **⚠️ GOTCHA for tooling authors:** this is still a *negotiated* transition. Mixed networks\n> run tor1 and CGO side by side until client\u002Frelay adoption saturates. When reading packet\n> traces or writing parsers, expect **both formats on the wire** through at least the rest of\n> 2026 — a parser that assumes the 4-byte SHA-1 digest, or one that assumes the 16-byte tag,\n> will silently mis-frame half the network.\n\n---\n\n## §5. Flow control, congestion control, and Conflux\n\nTor's performance problems were never mainly cryptographic; they're **queueing**. Three\nsuccessive systems matter.\n\n### §5.1 Classical flow control: SENDME windows\n\nThe original design gives each stream a 50-cell window and each circuit a 100-cell window;\nreceivers return a `SENDME` acknowledgment every N cells to refill the sender's credit. This\nprevents memory overload at exits but says nothing about **congestion inside the network** —\nqueues built up at slow relays, and the result was Tor's infamous multi-second latency spikes.\n\n### §5.2 Proposal 324: real congestion control (C tor 0.4.7, 2022; tuned through 0.4.8)\n\nProp 324 adapts TCP-grade algorithms to circuits: **Tor-Westwood** and **Tor-Vegas**\nestimators pace cells by measured RTT and bandwidth-delay product, with **NOLA** as a\nbandwidth-oracle fallback. Circuit windows can now grow and shrink, and new round-trip\nsignaling (`XON`\u002F`XOFF`-style flow control cells on the newer subprotocol) replaces some\nSENDME behavior. Measured effect: large-file and interactive performance improved markedly,\nwith much smaller inter-relay queues. In Arti, flow control + congestion control\n(`flowctl-cc`) became **stable in 2.4.0 (June 2026)** and always-on with CGO in 2.6.x.\n\n### §5.3 Proposal 329: Conflux (traffic splitting)\n\nConflux lets a client split one logical stream across **two circuits to the same exit**,\nreassembling out-of-order data at the endpoints — head-of-line blocking on one circuit no\nlonger stalls everything. The relay side adds `RELAY_CONFLUX_LINK`\u002F`LINKED`\u002F`SWITCH` cells and\nreorder buffers; scheduling algorithms (MinRTT, LowRTT, BLEST) decide which leg gets the next\ncell. Status: **enabled on the live network in C tor since 2024** (controlled by the\n`cfx_enabled` consensus parameter), with bug-fix work through 2025–2026 (a non-fatal assertion\nfixed in 0.4.8.20; a Conflux queue DoS, CVE-2026-44600, fixed in 0.4.9.7). Benefits are\nlargest for interactive\u002Fbursty traffic. Arti porting was active through 2025 (Arti 1.4.5\nchangelog) — check release notes for client-side status.\n\n> **⚠️ GOTCHA for onion-service operators:** the legacy Python `vanguards` addon is\n> **incompatible with Conflux** (it requires `ConfluxEnable 0`). That is one of the main\n> reasons the vanguards path now runs through Arti — see §9.1 →\n> `tor-onion-services-and-hardening`.\n\n---\n\n## Where to go next\n\n- The relays, the nine directory authorities, consensus mechanics and guard pinning:\n  §6–§7 → `tor-network-consensus-guards-and-paths`.\n- v3 onion services, vanguards and proof-of-work: §8–§9 →\n  `tor-onion-services-and-hardening`.\n- Bridges, pluggable transports and the 2025–2026 censorship arms race: §10 →\n  `tor-censorship-circumvention-and-bridges`.\n- What actually deanonymizes people: §11 → `tor-attack-literature-and-threat-model`.\n- Specs, proposals and the foundational papers: §16 →\n  `tor-ecosystem-alternatives-and-governance`.\n","# Tor: What It Promises, How Circuits Are Built, and What Encrypts a Cell\n\n> **Part 1 of 8** of the *Tor and Onion Networks* reference (plugin `tor-and-onion-networks`),\n> covering §1–§5. Sibling skills: `tor-network-consensus-guards-and-paths` (§6–§7),\n> `tor-onion-services-and-hardening` (§8–§9),\n> `tor-censorship-circumvention-and-bridges` (§10),\n> `tor-attack-literature-and-threat-model` (§11),\n> `tor-ecosystem-alternatives-and-governance` (§12, §14–§16),\n> `tor-building-on-tor-software-patterns` (§13.1–§13.4, handbook §1),\n> `tor-running-relays-bridges-and-hardware` (§13.5, handbook §2–§6).\n> Section numbers are shared across the set; a reference written as §N → `skill` points into\n> that sibling skill.\n>\n> **Currency:** compiled 16 September 2026. Protocol fundamentals (cell formats, handshake\n> math) are stable and change slowly. **Version numbers and deployment status are dated** —\n> the tor1 → CGO transition in §4 and the Conflux rollout in §5 are mid-flight, so verify\n> against \u003Chttps:\u002F\u002Fspec.torproject.org\u002F> before relying on them months from now.\n\n> **⚠️ A low-latency anonymity overlay: three independently operated relays and layered\n> encryption, so no single party links who you are to what you are doing.**\n>\n> **⚠️ GOTCHA** boxes mark where the mental model people carry is wrong in ways that cause\n> real deanonymization or broken tooling.\n>\n> **The three ideas that organize this document:**\n> 1. **⚠️ TOR SEPARATES IDENTIFICATION FROM ROUTING** (§1, §3). **The guard knows who you\n>    are but not what you do; the exit knows what you do but not who you are. Every design\n>    decision downstream — guards, families, path restrictions — exists to keep those two\n>    facts in different hands.**\n> 2. **⚠️ THE TLS BETWEEN RELAYS IS NOT THE ANONYMITY LAYER** (§3.3). **Hop-to-hop TLS\n>    protects against someone sitting between relays. The anonymity comes from the\n>    onion-encrypted relay cells riding inside it. Confusing the two is the most common\n>    architectural misreading of Tor.**\n> 3. **⚠️ TOR'S PERFORMANCE PROBLEM WAS NEVER CRYPTOGRAPHIC — IT WAS QUEUEING** (§5).\n>    **Twenty years of latency complaints came from unmanaged queues at slow relays, which\n>    is why the 2022–2026 work is congestion control and traffic splitting, not faster\n>    ciphers.**\n\n---\n\n## §1. What Tor is, what it promises, and what it does not\n\nTor is a low-latency anonymity overlay network: roughly 9,200 volunteer relays and 2,300\nbridges carrying traffic for several million daily users (network-size and user CSV data\nfrom [Tor Metrics](https:\u002F\u002Fmetrics.torproject.org\u002F), 8–14 Sept 2026; see §6 →\n`tor-network-consensus-guards-and-paths` for detail). A client's traffic is routed through a\nchain (\"circuit\") of three independently operated relays, with layered (\"onion\") encryption\nsuch that:\n\n- the **first relay (guard)** knows *who you are* but not what you're doing;\n- the **middle relay** knows neither;\n- the **exit relay** (or onion-service rendezvous point) knows *what is being done* but not\n  *by whom*.\n\nThe design goal is **unlinkability**: no single entity — relay, ISP, destination server —\nshould be able to connect a user's identity to their activity. Tor provides *transport-layer*\nanonymity: it hides network endpoints, not the content or the behavior inside the stream.\n\n### ⚠️ What Tor deliberately does not promise\n\nPer its own threat model:\n\n- **Protection against a global passive adversary.** Anyone who can watch both ends of a\n  circuit (e.g., your ISP *and* the destination's network) can confirm relationships by\n  timing\u002Fvolume correlation. Tor is low-latency by design; defeating end-to-end correlation\n  is what high-latency mix networks try to do instead (§14 →\n  `tor-ecosystem-alternatives-and-governance`).\n- **Confidentiality past the exit relay.** Exit traffic to a clearnet site leaves the Tor\n  envelope. Use end-to-end encryption (HTTPS) as usual; Tor Browser bundles HTTPS-only mode\n  partly for this reason.\n- **Application-level anonymity.** Logging into your personal account over Tor, or running a\n  browser that leaks identifying fingerprints, defeats the network layer. This is why the Tor\n  Project ships a hardened browser rather than telling people to point any app at a SOCKS\n  proxy.\n- **Invisibility.** A censor can see that you *are using* Tor unless you use a pluggable\n  transport (§10 → `tor-censorship-circumvention-and-bridges`). Tor hides destinations, not\n  the fact of its use.\n\nA useful mental model: Tor gives you **three organizational trust boundaries** per circuit\nplus layered crypto, and then throws engineering effort (guards, congestion control, padding\nresearch, anti-enumeration transports) at shrinking the residual deanonymization surface.\n\n---\n\n## §2. History and lineage\n\n| Era | Development |\n|---|---|\n| Mid-1990s | Onion routing invented at the U.S. Naval Research Laboratory (Paul Syverson, David Goldschlag, Michael Reed). The insight: separate *identification* from *routing* by nesting encrypted hops. |\n| 2002 | First-generation onion routing prototype deployed at NRL. |\n| 2003–2004 | Roger Dingledine, Nick Mathewson, and Paul Syverson build the second-generation system — \"Tor\" — fixing the first gen's problems (no directory system, replay\u002Fintegrity weaknesses, usability). The USENIX Security 2004 paper *Tor: The Second-Generation Onion Router* remains the canonical design document. |\n| 2004–2006 | EFF funds early development; the network opens to the public. |\n| Dec 2006 | The Tor Project, Inc. founded as a U.S. 501(c)(3) nonprofit. |\n| 2008–2011 | Tor Browser Bundle (originally \"Torbutton\") makes the system usable by non-experts; bridges introduced to counter blocking in Iran and China. |\n| 2013–2014 | Snowden-era scrutiny and usage spike; ntor replaces the old RSA\u002FTAP handshakes for circuit extension; hidden-service deanonymization research (and at least one real-world attack campaign that surfaced in 2014) catalyzes the v3 redesign. |\n| 2015–2018 | `.onion` registered as a special-use domain (RFC 7686); v3 onion service spec written (224\u002F225 proposals era); ed25519 relay identities roll out. |\n| 2017 | Directory authority diversity pushed; Tor Browser 7 based on Firefox ESR 52 brings multiprocess. |\n| Oct 2021 | v2 onion services (16-char, RSA-1024, SHA-1-derived) **removed from the network**. All onion services are now v3 (56-char). |\n| 2022–2023 | Congestion control (proposal 324) ships in C tor 0.4.7 and is tuned through 0.4.8; Conflux traffic-splitting (proposal 329) developed. |\n| 2023 | Onion-service proof-of-work DoS defense (proposal 327) ships in tor 0.4.8 (Aug 2023). |\n| 2024 | Conflux enabled network-wide in C tor (controlled by consensus parameter); Arti (the Rust rewrite) reaches **client-side feature parity** and gets native vanguards. Tor Project and Tails merge organizations. WebTunnel transport introduced. |\n| 2025 | Tor 0.4.9 alphas introduce **CGO cell crypto** and **happy families**; Tor Browser 15 (Firefox ESR 140); Arti 2.x series begins; Tor VPN beta (Arti-based Android VPN); Conflux hardening continues; Tails 7 (Debian 13) and Qubes 4.3 ship. |\n| 2026 (as of Sept) | C tor 0.4.9.5 is the first stable 0.4.9 (12 Feb 2026); the 0.4.8 series is on the deprecation path. Arti 2.6.0 (1 Sept 2026) makes CGO and congestion control always-on, and pushes toward **Arti-as-relay** (directory mirror, DNS streams, channel auth). |\n\n**⚠️ The pattern to internalize:** **C tor is in managed decline, Arti is the future.** In\n2026 the frontier work is relay\u002Fdirectory-authority support in Arti + post-SHA-1 cell crypto\n(CGO) + congestion\u002Fflow-control modernization.\n\n---\n\n## §3. The core protocol: cells, circuits, and the telescoping handshake\n\nEverything in Tor happens in **cells** — fixed-size messages multiplexed over TLS connections\n(\"channels\") between relays. The classic cell is 514 bytes on the wire: a 2-byte circuit ID\n(4 bytes on modern link protocols), a 1-byte command, and a 509-byte payload. Variable-length\ncells exist for protocol negotiation and directory data.\n\nCell commands split into two families:\n\n- **Channel-level:** `CREATE`\u002F`CREATED` (now `CREATE2`\u002F`CREATED2`), `DESTROY`, `NETINFO`,\n  `CERTS`, `AUTH_CHALLENGE`\u002F`AUTHENTICATE`, `PADDING` — negotiated directly with the\n  immediate neighbor.\n- **Circuit-level:** `RELAY` (and `RELAY_EARLY` near circuit creation) — onion-encrypted\n  payloads that tunnel *through* the channel, understood only by whichever hop peels that\n  layer.\n\n### §3.1 How a circuit is built (\"telescoping\")\n\n1. **Client → Guard.** The client performs an authenticated ntor key exchange (`CREATE2`)\n   with its chosen guard. Both sides now share a symmetric key pair (forward\u002Fbackward\n   digests) for that hop. Result: circuit hop 1.\n2. **Extend to hop 2.** The client builds an `EXTEND2` relay cell containing the link\n   specifiers of the middle relay plus a fresh ntor handshake blob, onion-encrypted for the\n   guard. The guard unwraps, opens (or reuses) a channel to the middle, forwards the\n   handshake, and returns the middle's `CREATED2` response inside an `EXTENDED2` relay cell.\n   The client now has shared keys with hop 2 as well — negotiated *through* hop 1 without hop\n   1 learning the hop-2 key.\n3. **Extend to hop 3 (exit).** Same procedure, now layered through hops 1 and 2.\n\nEach hop strips one layer of symmetric crypto from outbound cells and adds one layer to\ninbound cells — the \"onion.\" The exit decrypts the final layer and forwards plaintext to the\ndestination (for client circuits) or the service.\n\n```\n\u002F\u002F Outbound (client → destination)\nClient builds:  Layer1(Layer2(Layer3(payload)))\nGuard:  strips Layer1 → forwards Layer2(Layer3(payload)) to middle\nMiddle: strips Layer2 → forwards Layer3(payload)         to exit\nExit:   strips Layer3 → forwards payload                 to destination\n\n\u002F\u002F Inbound (destination → client) reverses: exit adds Layer3, middle adds Layer2,\n\u002F\u002F guard adds Layer1, and the client — holding keys for all three hops — strips all three.\n```\n\nThis is the whole reason the client must hold keys for every hop and each hop holds keys only\nfor itself: the client is the only party that can construct or unwrap the full stack.\n\nTwo safety rails worth knowing:\n\n- **`RELAY_EARLY` cells** bound circuit construction: only the first N (8) relay cells in a\n  circuit's life may extend it, which prevents infinitely long circuits being used as a\n  resource-exhaustion vector and bounds certain tagging\u002Fconfirmation games.\n- Circuit IDs scope cells within a channel; `NETINFO` cells exchange timing\u002Faddress data used\n  for NAT traversal and address confirmation.\n\n### §3.2 Streams\n\nOnce a circuit exists, the client attaches **streams** (TCP connections) to it with\n`RELAY BEGIN`; the exit answers `CONNECTED`. Data flows as `RELAY DATA` cells; `RELAY END`\ncloses a stream. Tor multiplexes many streams over one circuit, and one client's circuits may\ncarry arbitrarily many streams — but see **stream isolation** (§13.1 →\n`tor-building-on-tor-software-patterns`) for why Tor Browser splits streams by first-party domain\nanyway.\n\n### §3.3 ⚠️ GOTCHA: where the TLS ends\n\nA subtlety newcomers miss: the hop-to-hop transport is TLS over TCP, but that TLS only\nprotects against attackers sitting *between* relays — it is **not** the anonymity layer. The\nanonymity comes from the onion-encrypted relay cells riding inside. Relay channels\nauthenticate with ed25519 identities (since the 0.3.0 era), and as of C tor 0.4.9 (Feb 2026)\nthe last pre-ed25519 link authentication methods (v1\u002Fv2 handshakes, the old\n`RSA-SHA256-TLSSecret` auth) were **removed**, along with fake-TLS cipher advertisement code.\n\n---\n\n## §4. Circuit cryptography: from TAP to ntor to CGO\n\n### §4.1 The handshake stack\n\n| Generation | Used for | Status (Sept 2026) |\n|---|---|---|\n| TAP | RSA-1024 + DH-1024 first-hop create, old extend | **Fully removed** from C tor in 0.4.9 (2026) |\n| ntor | Curve25519\u002FX25519 one-way-authenticated key exchange for `CREATE2`\u002F`EXTEND2`; variant `hs-ntor` for onion-service intro\u002Frendezvous handshakes | Standard, universal |\n| **ntor-v3** | Extended ntor with authenticated client→server data in the handshake (needed by CGO's crypto negotiation) | Rolling out; required for CGO negotiation |\n\nThe ntor design (Goldberg–Stebila–Ustaoglu, 2012–2013) gives **one-way authentication in a\nsingle round trip**: the server proves identity via its long-term key, the client stays\nanonymous, and both derive forward-secret session keys. The cost that motivated it — the old\nTAP handshake needed full RSA operations per hop and was CPU-dominant for relays — is why\nmodern Tor scales at all.\n\n### §4.2 The cell crypto itself: \"tor1\" and its replacement\n\nFor two decades the relay-level symmetric crypto (\"tor1\") was AES-128 in counter mode plus a\n4-byte running SHA-1 digest for integrity. Cryptographers flagged its problems for years:\n\n- **Tagging attacks.** CTR mode is malleable: a hop (or observer who controls a hop) can flip\n  predictable bits in a cell and watch the effect emerge at another hop — a marked cell\n  becomes a tracking beacon through the circuit.\n- **Weak integrity.** Four bytes of digest is ~32-bit authentication, and the underlying hash\n  was SHA-1.\n- **No forward secrecy at the cell layer** — keys stayed constant for the life of the circuit.\n\nThe replacement, **Counter Galois Onion (CGO)**, designed by Degabriele, Melloni, Münch and\nStam (proposal 308, 2019; revised and instantiated as **proposal 359, \"CGO Redux\"**), is the\nbiggest change to Tor's wire format since v3 onion services:\n\n- Built on **UIV+**, a *rugged pseudorandom permutation*: encryption is non-malleable (any\n  tamper, and that cell *plus every subsequent cell in that direction* becomes unrecoverable\n  — killing tagging), while relays process cells via a deliberately malleable decrypt path.\n- Primitives are AES-128 (counter PRF) + **POLYVAL** universal hashing — chosen to exploit\n  AES-NI and PCLMULQDQ hardware instructions, so the security upgrade is also fast on modern\n  server CPUs.\n- Cell format becomes a **16-byte authenticator tag + 493-byte body**; the running SHA-1\n  digest and the `Recognized` field disappear (SENDMEs now bind to the cell tag).\n- Keys **evolve per cell** (`UPDATE_UIV`), so old cells can't be re-decrypted later even with\n  the current key — forward secrecy at the cell layer.\n- **Deployment status (Sept 2026):** implemented in C tor 0.4.9 (clients and relays negotiate\n  it) and in Arti, where CGO + congestion control became **always-on in Arti 2.6.0 (1 Sept\n  2026)**, after being marked stable in Arti 2.5.0 (30 June 2026). It requires ntor-v3 and the\n  new FlowCtrl congestion signaling — which is why it arrived together with the flow-control\n  redesign below. Onion-service-circuit CGO was in experimental stabilization in Arti 2.5.x\n  (mid-2026).\n\n> **⚠️ GOTCHA for tooling authors:** this is still a *negotiated* transition. Mixed networks\n> run tor1 and CGO side by side until client\u002Frelay adoption saturates. When reading packet\n> traces or writing parsers, expect **both formats on the wire** through at least the rest of\n> 2026 — a parser that assumes the 4-byte SHA-1 digest, or one that assumes the 16-byte tag,\n> will silently mis-frame half the network.\n\n---\n\n## §5. Flow control, congestion control, and Conflux\n\nTor's performance problems were never mainly cryptographic; they're **queueing**. Three\nsuccessive systems matter.\n\n### §5.1 Classical flow control: SENDME windows\n\nThe original design gives each stream a 50-cell window and each circuit a 100-cell window;\nreceivers return a `SENDME` acknowledgment every N cells to refill the sender's credit. This\nprevents memory overload at exits but says nothing about **congestion inside the network** —\nqueues built up at slow relays, and the result was Tor's infamous multi-second latency spikes.\n\n### §5.2 Proposal 324: real congestion control (C tor 0.4.7, 2022; tuned through 0.4.8)\n\nProp 324 adapts TCP-grade algorithms to circuits: **Tor-Westwood** and **Tor-Vegas**\nestimators pace cells by measured RTT and bandwidth-delay product, with **NOLA** as a\nbandwidth-oracle fallback. Circuit windows can now grow and shrink, and new round-trip\nsignaling (`XON`\u002F`XOFF`-style flow control cells on the newer subprotocol) replaces some\nSENDME behavior. Measured effect: large-file and interactive performance improved markedly,\nwith much smaller inter-relay queues. In Arti, flow control + congestion control\n(`flowctl-cc`) became **stable in 2.4.0 (June 2026)** and always-on with CGO in 2.6.x.\n\n### §5.3 Proposal 329: Conflux (traffic splitting)\n\nConflux lets a client split one logical stream across **two circuits to the same exit**,\nreassembling out-of-order data at the endpoints — head-of-line blocking on one circuit no\nlonger stalls everything. The relay side adds `RELAY_CONFLUX_LINK`\u002F`LINKED`\u002F`SWITCH` cells and\nreorder buffers; scheduling algorithms (MinRTT, LowRTT, BLEST) decide which leg gets the next\ncell. Status: **enabled on the live network in C tor since 2024** (controlled by the\n`cfx_enabled` consensus parameter), with bug-fix work through 2025–2026 (a non-fatal assertion\nfixed in 0.4.8.20; a Conflux queue DoS, CVE-2026-44600, fixed in 0.4.9.7). Benefits are\nlargest for interactive\u002Fbursty traffic. Arti porting was active through 2025 (Arti 1.4.5\nchangelog) — check release notes for client-side status.\n\n> **⚠️ GOTCHA for onion-service operators:** the legacy Python `vanguards` addon is\n> **incompatible with Conflux** (it requires `ConfluxEnable 0`). That is one of the main\n> reasons the vanguards path now runs through Arti — see §9.1 →\n> `tor-onion-services-and-hardening`.\n\n---\n\n## Where to go next\n\n- The relays, the nine directory authorities, consensus mechanics and guard pinning:\n  §6–§7 → `tor-network-consensus-guards-and-paths`.\n- v3 onion services, vanguards and proof-of-work: §8–§9 →\n  `tor-onion-services-and-hardening`.\n- Bridges, pluggable transports and the 2025–2026 censorship arms race: §10 →\n  `tor-censorship-circumvention-and-bridges`.\n- What actually deanonymizes people: §11 → `tor-attack-literature-and-threat-model`.\n- Specs, proposals and the foundational papers: §16 →\n  `tor-ecosystem-alternatives-and-governance`.\n",{"name":5,"description":7},"develop","ae32736a1aabddabca5d2c96d665aa0ebc7d8700","src\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Ftor-and-onion-networks\u002Fskills\u002Ftor-protocol-circuits-and-cell-cryptography",3,"https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey","https:\u002F\u002Fthe-vibey-project.github.io\u002Fvibey\u002F","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project","https:\u002F\u002Fvibewithadam.matthewsteinberger.com","freedom for software engineers","2026-11-15 05:47:42",{"source":8,"skillName":5,"commands":35,"prompt":45,"zipAvailable":46},[36,39,42],{"tool":37,"command":38},"npx","npx skills add the-vibey-project\u002Fvibey --skill tor-protocol-circuits-and-cell-cryptography",{"tool":40,"command":41},"bunx","bunx skills add the-vibey-project\u002Fvibey --skill tor-protocol-circuits-and-cell-cryptography",{"tool":43,"command":44},"pnpm","pnpm dlx skills add the-vibey-project\u002Fvibey --skill tor-protocol-circuits-and-cell-cryptography","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Ftor-and-onion-networks\u002Fskills\u002Ftor-protocol-circuits-and-cell-cryptography\u002FSKILL.md",true,[],[49,56,63,70,77,84],{"skillNo":50,"name":51,"nameZh":6,"description":52,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":53,"categories":54,"keywords":55,"publishedAt":19},"01M7AT1FQJAE1FQYAMDTF490Y4","white-paper-writing","Use when writing white papers, technical reports, policy documents, thought leadership pieces, or any long-form authoritative document. Triggers on requests to write, structure, outline, or review white papers, technical reports, position papers, policy briefs, or evidence-driven arguments. Also use for B2B content marketing strategy, technical documentation requiring authority, academic-style reports, or any document that must educate while persuading. Use when someone asks \"how should I structure this\" for a long-form document.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fwhite-paper-writing\u002FSKILL.md",[],[],{"skillNo":57,"name":58,"nameZh":6,"description":59,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":60,"categories":61,"keywords":62,"publishedAt":19},"01M7AT16BJRJJKK2MASXRZS76G","prose-technical-mechanics","Use when editing or writing prose at a technical\u002Fmechanical level: choosing between Anglo-Saxon and Latinate diction, placing information in topic vs stress positions, writing periodic vs cumulative sentences, handling sentence rhythm, punctuation as logic, transitions, paragraph structure, or phonesthetics. Use when someone wants to understand WHY prose works or doesn't, not just surface-level style advice. Triggers on word choice, sentence structure, emphasis, rhythm, diction, or any deep writing mechanics question.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fprose-technical-mechanics\u002FSKILL.md",[],[],{"skillNo":64,"name":65,"nameZh":6,"description":66,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":67,"categories":68,"keywords":69,"publishedAt":19},"01M7AT0W9HT0ZZ9Y287NW99MR3","prose-mechanics","Use when writing, editing, or analyzing prose style; when trying to avoid generic AI-sounding writing; when applying specific prose styles such as journalistic, academic, narrative, KJV biblical, Hemingway plain, Latinate\u002FJamesian, Orwellian, Augustan, Victorian, Early Modern (Bacon, Browne, Milton), or free indirect discourse; or when coaching someone on writing voice and mechanics. Triggers on any request to improve, analyze, or write in a specific prose style, to fix \"AI-sounding\" writing, to make prose more vivid or concrete, to vary sentence rhythm, or to understand how a particular writer achieves their effects.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fprose-mechanics\u002FSKILL.md",[],[],{"skillNo":71,"name":72,"nameZh":6,"description":73,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":74,"categories":75,"keywords":76,"publishedAt":19},"01M7AT0KR635JX66X2X6F3E2CP","narrative-structure","Use when structuring stories, pitches, presentations, white papers, or any persuasive narrative; analyzing the hero's journey in fiction or non-fiction; applying StoryBrand for business communication; designing narrative arcs for thought leadership; or understanding Vogler, Snyder, Propp, or Campbell frameworks. Triggers on any narrative structure, storytelling, pitch structure, StoryBrand, hero's journey, three-act structure, Freytag, Save the Cat, narrative arc, or \"how should I structure this story\u002Fpitch\u002Fpresentation\" request.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fnarrative-structure\u002FSKILL.md",[],[],{"skillNo":78,"name":79,"nameZh":6,"description":80,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":81,"categories":82,"keywords":83,"publishedAt":19},"01M7AT0AKGW09NKKZDYBT8VRDE","legal-brief-writing","Use when writing legal briefs, motions, or any court filing; when analyzing judicial cognition and persuasion; when applying elite brief-writing techniques from Roberts, Clement, Kagan, Garner; or when structuring legal arguments for maximum persuasion. Also use for autistic\u002Fneurodivergent advocates who need to understand how their natural communication style maps to elite legal advocacy. Triggers on any legal writing, brief drafting, motion preparation, or legal argument structuring request.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Flegal-brief-writing\u002FSKILL.md",[],[],{"skillNo":85,"name":86,"nameZh":6,"description":87,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":88,"categories":89,"keywords":90,"publishedAt":19},"01M7AT00VWPS4Q8B5VEMXX9NQT","journal-publishability-criteria","Use whenever a research paper, article or preprint is being written, revised, reviewed or published, and specifically before every publication of docs\u002Fpaper.md. Covers what legitimate journals actually gate on, in order of gating — the editor's desk screen (scope, novelty, package problems), what reviewers judge (methodological soundness, claims the evidence can carry, statistical and qualitative reporting), EQUATOR reporting guidelines, the ICMJE authorship criteria, COPE's view of text recycling, the near-universal generative-AI policy (no AI author, disclosure, author responsibility), TOP 2025 transparency standards, how to verify a journal is legitimate, submitting without an institution, and a pre-submission checklist. Triggers on requests to evaluate, review, prepare or submit a paper, to judge publishability, to write a declarations or data-availability section, to disclose AI use, or to check a journal.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fjournal-publishability-criteria\u002FSKILL.md",[],[]]