openclaw-crm

Local-first CRM for managing leads, deals, follow-ups, and pipelines via CLI using SQLite with WAL mode.

Before installing: 1) Inspect the interchange behavior — src/interchange.js writes state/deals .md files that include contact names, emails and deal values. If your workspace/interchange directory is shared with other agents or users, this will expose PII and monetary data. If you do not want that, modify or disable the refresh/write functions. 2) Verify the referenced writeMd import (../../interchange/src/index.js) exists in your environment or replace it with a local writer; the relative import looks like it will fail unless an interchange package is present at that path. 3) Run npm install in an isolated environment or container: better-sqlite3 builds native modules and may require build tools. 4) Audit dependencies (npm audit) and review package-lock.json for unexpected packages. 5) Be aware backups copy .db, .wal, .shm files — secure those files and their destinations. 6) If you plan to let agents run this autonomously, explicitly confirm which interchange files will be published and who/what can read workspace/interchange; the code currently writes sensitive state data which could be picked up by other agents. If you want, I can point to the specific lines that produce interchange/deal content and suggest small code edits to redact sensitive fields.

Type: OpenClaw Skill Name: openclaw-crm Version: 1.0.0 The skill is classified as suspicious due to a critical vulnerability in its `backup` command, located in `src/cli.js` and `src/backup.js`. The `--output <path>` option for `crm backup` allows an attacker (via prompt injection to the agent) to specify an arbitrary absolute file path. This enables the skill to write the database backup files (`.db`, `.db-wal`, `.db-shm`) to any location on the filesystem, potentially overwriting sensitive system files (e.g., `/etc/passwd.db`, `/etc/cron.d/malicious_job.db`). While the skill itself does not exhibit malicious intent, this arbitrary file write capability represents a significant security flaw that could be exploited for system compromise.