Summary
yoloshii/ClawMem v2026-03-20 was classified as CRITICAL RISK with a risk score of 1474. Sigil detected 92 findings across 123 files, covering phases including provenance, install hooks, network exfiltration, code patterns. Review the findings below before installing this package.
Package description: On-device context engine and memory for AI agents. Claude Code and OpenClaw. Hooks + MCP server + hybrid RAG search.
v2026-03-20
21 March 2026, 00:45 UTC
by Sigil Bot
Risk Score
1474
Findings
92
Files Scanned
123
Provenance
Repository
https://github.com/yoloshii/ClawMemScanned From
https://github.com/yoloshii/ClawMem.gitFindings by Phase
Phase Ordering
Phases are ordered by criticality, with the most dangerous at the top. Click any phase header to expand or collapse its findings. Critical phases are expanded by default.
install-pip-setup-exec
CRITICALsetup.py executes code at install time
repo/AGENTS.md:41
### Server Setup (all three use llama-server)
Why was this flagged?
This setup.py calls subprocess, os.system, exec, or eval during package installation. Legitimate packages rarely need to execute arbitrary commands at install time. This pattern is commonly used by malicious packages to download and run payloads, exfiltrate environment variables, or establish persistence. Rated CRITICAL because it runs with the installer's full permissions.
install-pip-setup-exec
CRITICALsetup.py executes code at install time
repo/CLAUDE.md:41
### Server Setup (all three use llama-server)
Why was this flagged?
This setup.py calls subprocess, os.system, exec, or eval during package installation. Legitimate packages rarely need to execute arbitrary commands at install time. This pattern is commonly used by malicious packages to download and run payloads, exfiltrate environment variables, or establish persistence. Rated CRITICAL because it runs with the installer's full permissions.
install-makefile-curl
HIGHMakefile/script pipes remote content to shell
repo/README.md:66
- [Bun](https://bun.sh) v1.0+ — runtime for ClawMem. On Linux, install via `curl -fsSL https://bun.sh/install | bash` (not snap — snap Bun cannot read stdin, which breaks hooks).
- SQLite with FTS5 — included with Bun. On macOS, install `brew install sqlite` for extension loading support (ClawMem detects and uses Homebrew SQLite automatically).Why was this flagged?
A script or Makefile pipes content from a remote URL directly into a shell (curl | sh or wget | bash). This is inherently dangerous because the remote content can change at any time, and the command runs with the current user's permissions. Rated HIGH because it requires manual execution (unlike install hooks) but still executes arbitrary remote code.
install-makefile-curl
HIGHMakefile/script pipes remote content to shell
repo/docs/quickstart.md:7
- [Bun](https://bun.sh) v1.0+ — install via `curl -fsSL https://bun.sh/install | bash`, not snap (snap Bun has stdin restrictions that break hooks)
- A GPU for local inference (default models need ~4GB VRAM; SOTA upgrade needs ~10GB). Or use [cloud embedding](guides/cloud-embedding.md)Why was this flagged?
A script or Makefile pipes content from a remote URL directly into a shell (curl | sh or wget | bash). This is inherently dangerous because the remote content can change at any time, and the command runs with the current user's permissions. Rated HIGH because it requires manual execution (unlike install hooks) but still executes arbitrary remote code.
install-makefile-curl
HIGHMakefile/script pipes remote content to shell
repo/docs/troubleshooting.md:9
- On Linux, Bun installed via snap (`/snap/bin/bun`) cannot read stdin due to snap's confinement sandbox. Hooks receive no prompt input and silently return empty context.
- Fix: Install Bun via the official installer (`curl -fsSL https://bun.sh/install | bash`) which places it at `~/.bun/bin/bun`. The `bin/clawmem` wrapper prefers `~/.bun/bin/bun` over the system bun for this reason. If hooks return empty on a snap-based system, verify `~/.bun/bin/bun` exists and is executable.
Why was this flagged?
A script or Makefile pipes content from a remote URL directly into a shell (curl | sh or wget | bash). This is inherently dangerous because the remote content can change at any time, and the command runs with the current user's permissions. Rated HIGH because it requires manual execution (unlike install hooks) but still executes arbitrary remote code.
install-pip-setup-exec
CRITICALsetup.py executes code at install time
repo/src/clawmem.ts:928
async function cmdSetup(args: string[]) {
const subCmd = args[0];Why was this flagged?
This setup.py calls subprocess, os.system, exec, or eval during package installation. Legitimate packages rarely need to execute arbitrary commands at install time. This pattern is commonly used by malicious packages to download and run payloads, exfiltrate environment variables, or establish persistence. Rated CRITICAL because it runs with the installer's full permissions.
Badge
Markdown
[](https://sigilsec.ai/scans/DD6522AE-742F-4E27-B8BA-BC57DE313F0A)HTML
<a href="https://sigilsec.ai/scans/DD6522AE-742F-4E27-B8BA-BC57DE313F0A"><img src="https://sigilsec.ai/badge/github/yoloshii/ClawMem" alt="Sigil Scan"></a>Run This Scan Yourself
Scan your own packages
Run Sigil locally to audit any package before it touches your codebase.
Early Access
Get cloud scanning, threat intel, and CI/CD integration.
Join 150+ developers on the waitlist.
Get threat intelligence and product updates
Security research, new threat signatures, and product updates. No spam.
Other github scans
Believe this result is incorrect? Request a review or see our Terms of Service and Methodology.