Is morluto/rea safe?
Yes. morluto/rea is safe to install. Oathe's behavioral security audit gave the rea skill by morluto a trust score of 86/100 with 4 findings, none critical or high.
https://github.com/morluto/rea
Is morluto/rea safe to install?
REA (rea-agents v6.4.0) is a legitimate, production-grade local reverse engineering tool with MCP/CLI interfaces. No malicious instructions, hidden payloads, prompt override attempts, or data exfiltration pathways were detected in the skill content or repository. The primary observations are expected network connections to GitHub for submodule resolution, a standard npm prepare lifecycle hook, and canary file reads attributable to the audit framework's own verification process rather than skill code execution.
What security issues were found in morluto/rea?
Category Scores
Findings (4)
LOW Submodule network connections to GitHub during audit -10 ▶
TCP connections to GitHub (140.82.121.3:443) were made. This is expected behavior for a repository with 5 declared git submodules. All submodule targets are legitimate open-source projects. No data encoding or covert channel behavior observed.
Evidence: 140.82.121.3:443 in unique destination IPs; .gitmodules with 5 entries; install failed before any code ran.
LOW Skill pre-authorizes analysis tool calls without per-call approval -7 ▶
SKILL.md instructs the agent that analysis tools can be called directly without per-call approval flags. While clarified as REA's internal authorization model (not overriding agent-level controls), this framing could cause an agent to execute analysis operations on sensitive targets without seeking user confirmation.
Evidence: SKILL.md: 'call available analysis tools directly. REA does not require permission grants or per-call approval flags.'
LOW npm prepare lifecycle hook -10 ▶
Standard npm lifecycle hook that would execute node scripts/prepare.mjs on npm install. Not triggered during this audit. Scripts appear to handle skill/metadata generation for release. Not malicious but represents code execution on install.
Evidence: package.json scripts.prepare = 'node scripts/prepare.mjs'
INFO Canary files read by audit framework during cleanup -10 ▶
Six sensitive canary files were opened and read at the end of the monitoring period, immediately after audit teardown began. Timing and process context is consistent with the oathe-audit system's own canary integrity verification. Files were not modified. No exfiltration path detected.
Evidence: inotify: .env, id_rsa, .aws/credentials accessed at 09:25:59; auditd: PATH records 22398-22403 at 1791797185.465 between audit teardown commands.
Should I install morluto/rea?
Oathe's verdict for morluto/rea is SAFE with a trust score of 86/100. Recommendation: Install.