@decoopus/chatgpt-setup@1.0.3
OpenSSF/OSV malicious advisory
published · by OpenSSF/OSVOpenSSF/OSV advisory MAL-2026-18053 confirms this npm version as malicious.
OpenSSF/OSV advisory MAL-2026-18053 confirms this npm version as malicious.
LPM's AI review found this version suspicious. That disagrees with the OpenSSF/OSV advisory, which supplies the block.
The CLI makes broad, risky changes to Codex permissions during an explicit setup command, which warrants a warning for agent capability abuse. No install-time trigger or separate credential exfiltration behavior was identified.
Where they are, most severe file first. Code opens a file's flagged lines, with their evidence, below.
package.json2 findingsbin/cli.js5 findingsThe manifest exposes a user-run setup command and defines only a start script, with no npm install hook.
package.jsonView on unpkg · L5The manifest exposes a user-run setup command and defines only a start script, with no npm install hook.
package.jsonView on unpkg · L15The CLI writes Codex configuration under CODEX_HOME or the user’s .codex directory.
bin/cli.jsView on unpkg · L183The CLI writes Codex configuration under CODEX_HOME or the user’s .codex directory.
bin/cli.jsView on unpkg · L178This report applies to @decoopus/chatgpt-setup@1.0.3.
See all versions of @decoopus/chatgpt-setup for other recorded verdicts.
Evidence last updated: .
Source advisory published: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.
Setup sets approval to never and the sandbox to full access.
bin/cli.jsView on unpkg · L237Setup changes persisted Codex permissions to full access and adds the home, Downloads, and system root to project access paths.
bin/cli.jsView on unpkg · L373Setup changes persisted Codex permissions to full access and adds the home, Downloads, and system root to project access paths.
bin/cli.jsView on unpkg · L380