Workcrm

A local-first CRM that drafts and requires explicit user confirmation before saving contact or record changes for auditability.

This package appears to be a legitimate local CRM implemented in Python and enforces an explicit confirm gate, but the runtime docs don't fully match the code. Before installing or enabling auto-invocation, consider: 1) The SKILL.md CLI (python -m workcrm) likely doesn't exist (no __main__ or console script); try importing WorkCRMEngine/WorkCRMRepo in a sandbox or running the tests instead. 2) The SKILL.md default DB path (~/.openclaw/workcrm/workcrm.sqlite3) differs from the code's default (~/.openclaw/data/workcrm/workcrm_work.sqlite) — confirm where the DB will be created. 3) SKILL.md mentions WORKCRM_DB_PATH, but the package/registry don't declare it — if you rely on an env override, test how the agent will pass it. 4) The skill will create directories and a SQLite DB under your home directory and retain drafts (including rejected drafts) — if that is sensitive, point the skill at a dedicated data directory. 5) There are no network calls or credential requests in the code, which reduces exfiltration risk. If you want to proceed: run the included tests or import the package in an isolated environment to validate behavior, or ask the publisher to fix the SKILL.md / packaging mismatches before enabling the skill for autonomous agent use.

Type: OpenClaw Skill Name: workcrm Version: 0.1.0-alpha.4 The OpenClaw WorkCRM skill is classified as benign. The code adheres to good security practices, utilizing parameterized queries in `workcrm/db/repo.py` to prevent SQL injection. Input parsing in `workcrm/parse.py` uses regular expressions to extract data, which is then passed as values to the database, not as executable code. The `SKILL.md` documentation provides clear usage instructions and does not contain any malicious prompt injection attempts. File system access is limited to its own SQLite database, located in a standard OpenClaw data directory, and there are no network calls or attempts to access sensitive system files. The explicit confirmation gate (`记`/`不记`) adds a layer of security by requiring user approval before committing actions.