Jmail World - Search Epstein Files, E-Mails & Messages
PassAudited by ClawScan on May 1, 2026.
Overview
This skill appears purpose-aligned for searching the jmail.world archive, but it uses local scripts, external network downloads, and a temporary cache that users should understand before use.
Before installing, understand that this skill is for querying jmail.world and may run local shell scripts, contact jmail.world/data.jmail.world, download potentially large public archive files, and cache Parquet data in /tmp. There is no evidence in the provided artifacts of credential theft, hidden persistence, or unrelated behavior.
Findings (3)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
If used, the agent may query jmail.world and download archive files to the local machine at the user’s direction.
The skill exposes shell-based search and download workflows. This is expected for archive research, but users should notice that invoking the skill can make external requests and write downloaded files.
Use `web_fetch` or `curl`... Use the included scripts: `bash scripts/jmail-search.sh "query text"` ... `bash scripts/jmail-duckdb.sh download "EFTA02406146"`
Use the commands only for intended archive research, review download destinations, and avoid running broad downloads unless you want those files saved locally.
The skill may fail or prompt manual installation of external tools before it can run DuckDB-based searches.
The helper script depends on DuckDB even though the registry requirement section declares no required binaries. The dependency is disclosed in the script and purpose-aligned, but metadata is incomplete.
if ! command -v duckdb &>/dev/null; then ... echo " apt install duckdb OR brew install duckdb"
Install only trusted versions of required tools such as DuckDB, curl, and python3, and treat the registry’s missing dependency declaration as something to verify manually.
Archive data can remain temporarily on the local system and be reused across skill invocations.
Downloaded Parquet datasets are cached locally and reused for later queries. This is efficient and purpose-aligned, but users should know results may come from locally cached archive data.
CACHE_DIR="/tmp/jmail-cache" ... Re-download if older than 24 hours or doesn't exist
Clear /tmp/jmail-cache if you do not want local cached copies retained, and be aware that cached data may affect later search results until refreshed.