Credit
The Independent 100 list was curated by dan (@irl_danB) and published on 2026-09-24 at independent.prose.md (announcement). The ranks and section names are his, and number 100 is his open wild-card slot. This study is ours and is not endorsed by him or by anyone on the list.
What they publicly do
Pseudonymous account. This entry records public metadata only, for defensive study: no repository contents were opened, run or summarised. The GitHub account elder-plinius, whose profile page names @elder_plinius, hosts widely starred AGPL-3.0 repositories, among them CL4R1T4S, whose README also names @elder_plinius [Verified: evidence/credits.jsonl]. A post from the account links bt6.gg, which describes itself as an independent frontier AI red team [Verified: X API read of x.com/elder_plinius/status/2099061881302614040, 2026-09-25, evidence/x-reads.jsonl].
- X profile: x.com/elder_plinius
- GitHub: github.com/elder-plinius
- Also: bt6.gg
Public work
- CL4R1T4S repo AGPL-3.0 last activity 2026-09-22 · 50,655 GitHub stars on 2026-09-24
Described as a collection of system prompts from AI products and coding tools [Verified: repository description only; contents not opened].
- BT6 other no license stated
Describes itself as an independent frontier AI red team [Reported: bt6.gg].
Licenses of GitHub repositories are from the authenticated GitHub license API on 2026-09-25, one row per repository in evidence/licenses.jsonl (“no license file” means the API found none; “license file present, not recognised by GitHub” means GitHub could not classify it). Other licenses are as the source states. Check the license file before reusing anything.
Relevance to our work
3 of 5 lenses: security, agents, mcp, evals
Defensive relevance only. The popularity and freshness of these repositories is a signal for defenders: agents and MCP servers should assume system prompts leak and that published jailbreak and injection techniques will reach them. The lesson is to keep no secrets or authority in prompts, to separate trusted control from untrusted data, and to measure injection resistance with published benchmarks. These repos are not for adoption.
Scored 0 to 5 against the study’s relevance rubric: 5 means public, openly licensed work squarely in our lanes that we could study or adopt now, and 1 means little connection to our work. The score is one reviewer’s judgement of how much this public work bears on our own work. It is not a ranking of people and says nothing about the person or the quality of their work.
What we would study next
Nothing is listed.
Before reusing anything: The repositories are AGPL-3.0 and describe themselves as attack material. Do not vendor them, run them or put them in agent context. No file contents were opened for this entry. Defensive study can use published research instead: AgentDojo (MIT), CaMeL (Apache-2.0 code) and the Constitutional Classifiers paper.
Sources
- github.com/elder-plinius
- github.com/elder-plinius/CL4R1T4S
- bt6.gg
- x.com/elder_plinius/status/2099061881302614040
Overall confidence in this record: Medium. Sources were read on 2026-09-24 and every link was checked on 2026-09-25. The full record is line 7 of people.jsonl.
Limits
Not verified: The purpose of the account's other repositories beyond their descriptions. The account's profile link and its posts other than the one cited were deliberately not opened. The CL4R1T4S README was fetched only to confirm that it names the handle.
This record comes from a first-pass survey of public material by an AI agent session on 2026-09-24, edited for publication on 2026-09-25. Nothing was installed or run for it. Roles, stars and licenses change; see the study’s limits.
Request a correction
Is something about this record wrong or out of date, or would you rather not be covered? Open a correction (a GitHub issue form with this page’s address filled in). A request to remove a record is honoured without argument; accepted corrections are recorded in the changelog with the date.