recess-cli@3.25.0
Safe Recess administration and family AI tools from the command line.
AI Security Review
scanned 3d ago · by lpm-firewall-aiThe published install automatically replaces a package-owned skill in global agent directories without asking first. This warrants a warning for agent-extension lifecycle risk, but the inspected behavior does not establish broader control-surface abuse or another concrete attack.
Decision evidence
public snapshot- package.json runs scripts/postinstall.mjs during installation.
- The postinstall hook imports and runs installSkill() for published installs, without a consent check.
- installSkill() targets the user's Codex and Claude skill directories, removes the existing recess-cli skill, and copies in the bundled skill.
- The apps init command writes AGENTS.md as part of an explicit project scaffolding action.
- The installer only refreshes this package's own named skill and removes one named superseded skill directory.
- The postinstall hook skips source checkouts and exits silently if the setup module is absent.
Source & flagged code
4 flagged · loading sourcePackage defines install-time lifecycle scripts.
package.jsonView on unpkgInstall-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgPackage contains a possible secret pattern.
dist/commands/onboarding.jsView on unpkg · L576Findings
Affected versions and remediation
This report applies to recess-cli@3.25.0.
See all versions of recess-cli for other recorded verdicts.
- Review the evidence and your use of recess-cli@3.25.0 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.