Cosmic
100% sUNC, 1.733, Detected. A GitBook documents the delivery path. A DevForum post documents a detection method.
A GitHub topic entry describes Swift as a "Level 8 Roblox executor" with 87% sUNC and 92% UNC. It adds a specific limitation: "frequent crashes limit its reliability, making it a neutral choice for users who can tolerate instability" .
That entry is a repository description, not a tracker page. It does not name a version number, a release date, or a source URL. It records a function score and a stability observation. The stability observation is unusual: most tracker entries do not mention crashes.
A community troubleshooting guide for Swift contains a step that deserves to be quoted directly: "Turn Off Your Antivirus. Temporarily disable antivirus — it can block or corrupt ScriptHub loading" .
The same guide instructs users to use a specific Roblox version hash (version-225e87fdb7254f64), to avoid bootstrappers that use FastFlags, and to "never use your main Roblox account" . It also includes a FAQ entry titled "Is Swift a Virus Is It Safe" with the answer "Swift is NOT a RAT False Positive Explanation" .
The guide is written by someone in the Swift community. It is not an independent audit. But the instruction to disable antivirus is a specific, documented practice. It is also a practice that security researchers consistently warn against, regardless of the software involved.
A ScriptBlox script page includes user comments about Swift. One user writes: "its not a logger i tried it i use swift works not op but what do you think for a free script" . Another replies that the script is "NOT a logger, it just doesn't support TRASH executors like velocity" .
These are user comments, not verification. The first comment asserts Swift is "not a logger" based on personal use. The second is dismissive of a different executor. Neither is a security audit. They are evidence that people use Swift and discuss it, not evidence that it is safe.
An OnWorks software listing for a different executor, Bunni, contains a sentence that connects it to Swift: "Originally rebranded as Swift, it has now been rebuilt and relaunched under the name Bunni Executor" . The same listing claims Bunni achieves "95% sUNC" .
That sentence implies Swift and Bunni are related by a rebrand. It is a single claim from a single listing. No independent source reviewed here confirms the relationship or the direction of the rebrand. It is recorded because it exists, not because it is verified.
Sources: GitHub topic entry · community troubleshooting guide · ScriptBlox comments · OnWorks listing . None of these is a security audit.
Most profiles on this site have a tracker entry from ScriptBunker, ScriptBlox, or Rscripts. Swift does not. No ScriptBunker page, ScriptBlox executor page, or Rscripts listing for Swift was found in the sources reviewed here. The 87% sUNC figure comes from a GitHub repository description , not from a tracker that runs tests and publishes results.
That distinction matters. A tracker entry is a measurement, even if third-party. A repository description is a claim. The 87% number may be accurate. It may be outdated. It may have been written by someone who ran the test once and never updated it. The record does not say.
The troubleshooting guide is the most specific source in this profile, because it documents what users are actually instructed to do . That instruction — disable antivirus, use a specific version hash, avoid the main account — is a description of the software's operational requirements. It is not a description of the software's safety.
One GitHub repository description says Swift is "completely keyless and secure from the start" . A ScriptBlox user comment also refers to Swift as a free executor that works without the complexity of keys . A keyless executor is one that does not require a key system, which is a distribution feature, not a security feature.
The same GitHub description uses the phrase "secure from the start" . That phrase is a claim, not a measurement. It is not supported by the troubleshooting guide, which instructs users to disable antivirus . A software that requires antivirus to be disabled before it runs is not "secure from the start" in any operational sense. The two statements are in tension. This page records both and resolves neither.
| Attribute | Value | Source | Confidence |
|---|---|---|---|
| sUNC | 87% | GitHub topic | Claim, not tracker |
| UNC | 92% | GitHub topic | Claim, not tracker |
| Stability | Frequent crashes | GitHub topic | Observation |
| Antivirus | Disable before launch | Troubleshooting guide | Instruction |
| Keyless | Claimed | GitHub repo | Claim |
| Official source | Not verified | — | Unknown |
| Tracker entry | Not found | — | Unknown |
"Claim" means the information comes from a repository description or user comment, not a tracker that publishes test results. "Unknown" means no source was found at review time.
No. It comes from a GitHub repository description , not a tracker page that runs tests and publishes results. No ScriptBunker, ScriptBlox, or Rscripts entry for Swift was found in the sources reviewed here.
The guide says antivirus "can block or corrupt ScriptHub loading" . That is an instruction from a community guide. It is not a security recommendation, and it is the opposite of what security researchers generally advise.
No. Keyless means the executor does not require a key system to use. It is a distribution feature, not a security feature. A keyless executor can still require antivirus to be disabled before it runs .
An OnWorks listing for Bunni says it was "originally rebranded as Swift" and later "rebuilt and relaunched under the name Bunni Executor" . That is a single claim from a single listing. No independent source confirms it.
Not yet. No verified official source was found. The Download button is a placeholder. When you have a confirmed source and choose to link it, replace the href. Until then, this page serves no file.
Because the evidence is thinner and more scattered. Cosmic had a tracker entry, a forum post, and a project manual. Swift has a repository description, a community troubleshooting guide, and user comments. The layout uses layers to separate those kinds of evidence, because they do not carry the same weight.
100% sUNC, 1.733, Detected. A GitBook documents the delivery path. A DevForum post documents a detection method.
100% sUNC, 99% UNC, free, Detected. Three tracker facts and several explicit unknowns.
ScriptBunker records 59% sUNC and Detected. A partial score is a compatibility measurement, not a safety rating.