ExperimentEXP-2026-0203v0.1

Rounds P1–P2 — individuals with AI-collaboration markers (400 repositories, 2026-10-13)

Public repositories owned by individual accounts, created within two years, 40 per round from each of CLAUDE.md, AGENTS.md, GEMINI.md, .github/copilot-instructions.md and .cursorrules. Surfaces: README, the drawn instruction file and agent-state files. The AI-collaboration profile (tree families, head-commit AI trailers, README declarations, star band) stayed sealed until label lock. A pattern seen in P1 was registered as a prediction before P2 was opened.

Research status: STABLE the current conclusions are relatively stable
Evidence level: E2 Controlled experiment
Result: POSITIVE
Data basis: BLIND PUBLIC DATA Real public files measured by an algorithm inside a sealed process. No person and no AI model read them; only coded features left the process. The numbers may be distorted.
Version: 0.1
Updated: 2026-10-13
Created: 2026-10-13
Domain: Privacy, Evaluation
Program: PRG-2026-0201 AI Agent Publication Boundary and Security Closure (APBSC)
Authors: Neo.K (EveMissLab)
AI collaborators: Splice (Claude Code, Anthropic) — Phase II protocols, runners, audits, reports and the integrated edition

Runs

run_count

metrics

rounds

P1

screened: 219
repositories: 200

P2

screened: 215
repositories: 200

repositories

400

Limitations

limitations: Exploratory, non-probability frame; not a prevalence estimate.
No human checked any finding; secret-like means a shape and was never tested.
Current public state only, observed once.

Reproduction

reproduction_instructions: python verify.py in the release. Per-repository profile rows are withheld (re-identification by rare combination), so S2's profile splits cannot be fully re-run from the release.

Results

Source	Relation	Target	Status	ID
`EXP-2026-0203` Rounds P1–P2 — individuals with AI-collaboration markers (400 repositories, 2026-10-13)	`produces`	`RST-2026-0203` Individuals: 2.75% publish .claude/settings.local.json, 10.5% an agent configuration file	`ACTIVE`	`REL-2026-0610`
`EXP-2026-0203` Rounds P1–P2 — individuals with AI-collaboration markers (400 repositories, 2026-10-13)	`produces`	`RST-2026-0204` Individual-era rule finds about twice the high-warning repositories of the organization-era rule (34 vs 17 of 400)	`ACTIVE`	`REL-2026-0615`
`EXP-2026-0203` Rounds P1–P2 — individuals with AI-collaboration markers (400 repositories, 2026-10-13)	`produces`	`RST-2026-0205` Registered prediction not supported: people using three or more AI tools do not leak more	`ACTIVE`	`REL-2026-0617`

Recorded fields

completed_at: 2026-10-13

Relations

Source	Relation	Target	Status	ID
`EXP-2026-0203` Rounds P1–P2 — individuals with AI-collaboration markers (400 repositories, 2026-10-13)	`runs_on`	`SYS-2026-0204` RBPM v0.9 runner — individuals with AI-collaboration markers (P1–P2)	`ACTIVE`	`REL-2026-0598`
`EXP-2026-0203` Rounds P1–P2 — individuals with AI-collaboration markers (400 repositories, 2026-10-13)	`produced`	`DAT-2026-0203` Phase II released features — individuals P1–P2	`ACTIVE`	`REL-2026-0599`
`EXP-2026-0203` Rounds P1–P2 — individuals with AI-collaboration markers (400 repositories, 2026-10-13)	`tests`	`THY-2026-0203` Agent context as an egress surface	`ACTIVE`	`REL-2026-0600`
`EXP-2026-0203` Rounds P1–P2 — individuals with AI-collaboration markers (400 repositories, 2026-10-13)	`tests`	`THY-2026-0204` From secret scanning to semantic privacy	`ACTIVE`	`REL-2026-0601`
`EXP-2026-0203` Rounds P1–P2 — individuals with AI-collaboration markers (400 repositories, 2026-10-13)	`tests`	`THY-2026-0207` Researcher-blind privacy measurement	`ACTIVE`	`REL-2026-0602`
`EXP-2026-0203` Rounds P1–P2 — individuals with AI-collaboration markers (400 repositories, 2026-10-13)	`produced`	`ART-2026-0223` P1 lock — protocol, instrument, runner and pre-lock digests `artifact://evemisslab/agent-publication-boundary/primary/P1/released/LOCK.json`	`ACTIVE`	`REL-2026-0603`
`EXP-2026-0203` Rounds P1–P2 — individuals with AI-collaboration markers (400 repositories, 2026-10-13)	`produced`	`ART-2026-0229` P1 pre-lock reading, written before group attributes were joined `artifact://evemisslab/agent-publication-boundary/primary/P1/PRELOCK_READING.md`	`ACTIVE`	`REL-2026-0604`
`EXP-2026-0203` Rounds P1–P2 — individuals with AI-collaboration markers (400 repositories, 2026-10-13)	`produced`	`ART-2026-0224` P2 lock — protocol, instrument, runner and pre-lock digests `artifact://evemisslab/agent-publication-boundary/primary/P2/released/LOCK.json`	`ACTIVE`	`REL-2026-0605`
`EXP-2026-0203` Rounds P1–P2 — individuals with AI-collaboration markers (400 repositories, 2026-10-13)	`produced`	`ART-2026-0230` P2 pre-lock reading, written before group attributes were joined `artifact://evemisslab/agent-publication-boundary/primary/P2/PRELOCK_READING.md`	`ACTIVE`	`REL-2026-0606`
`EXP-2026-0203` Rounds P1–P2 — individuals with AI-collaboration markers (400 repositories, 2026-10-13)	`produced`	`ART-2026-0212` Rounds P1–P2 report (individuals) `artifact://evemisslab/agent-publication-boundary/primary/V09_ROUNDS_REPORT.md`	`ACTIVE`	`REL-2026-0607`
`EXP-2026-0203` Rounds P1–P2 — individuals with AI-collaboration markers (400 repositories, 2026-10-13)	`produced`	`ART-2026-0214` S2 pooled report, individuals P1–P2 `artifact://evemisslab/agent-publication-boundary/secondary/S2-pool-individuals/REPORT.md`	`ACTIVE`	`REL-2026-0608`
`EXP-2026-0203` Rounds P1–P2 — individuals with AI-collaboration markers (400 repositories, 2026-10-13)	`produced`	`ART-2026-0217` S2 pooled outputs (JSON) `artifact://evemisslab/agent-publication-boundary/secondary/S2-pool-individuals/POOLED.json`	`ACTIVE`	`REL-2026-0609`
`EXP-2026-0203` Rounds P1–P2 — individuals with AI-collaboration markers (400 repositories, 2026-10-13)	`produces`	`RST-2026-0203` Individuals: 2.75% publish .claude/settings.local.json, 10.5% an agent configuration file	`ACTIVE`	`REL-2026-0610`
`EXP-2026-0203` Rounds P1–P2 — individuals with AI-collaboration markers (400 repositories, 2026-10-13)	`produces`	`RST-2026-0204` Individual-era rule finds about twice the high-warning repositories of the organization-era rule (34 vs 17 of 400)	`ACTIVE`	`REL-2026-0615`
`EXP-2026-0203` Rounds P1–P2 — individuals with AI-collaboration markers (400 repositories, 2026-10-13)	`produces`	`RST-2026-0205` Registered prediction not supported: people using three or more AI tools do not leak more	`ACTIVE`	`REL-2026-0617`

History and provenance

Canonical URL

https://evemisslab.com/ai/experiments/EXP-2026-0203/

Machine-readable

/ai/experiments/EXP-2026-0203/index.json

Snapshot

AI-SNAPSHOT-v0.1-58e716495670

Provenance

source: EveMissLab research collection: AI Agent Publication Boundary and Security Closure (真本體論14)
extracted_by: Splice (Claude Code), reading the open-source release's own manifests, locks and pooled outputs, and the hashes of the internal drafts
extracted_at: 2026-10-13
generator: tools/extract_all.py
claim_boundary: status, evidence level and result type follow the source artifact's own stated claim boundary; nothing is upgraded beyond what the report supports