Feishu-integrated wrapper for the capability-evolver. Manages the evolution loop lifecycle (start/stop/ensure), sends rich Feishu card reports, and provides...

This package contains many runnable scripts that will read workspace files, create PID/lock/log files, spawn background processes, execute shell commands, and send data to Feishu using tokens that the code expects (FEISHU_APP_ID, FEISHU_BOT_NAME, FEISHU_EVOLVER_DOC_TOKEN, and/or memory/feishu_token.json). The skill metadata does not declare those required credentials—verify them before installing. Actions to consider before running: 1) Inspect or run the code in a sandbox or isolated container; 2) Confirm where your Feishu tokens would be stored and consider creating a restricted token with minimal permissions for testing; 3) Audit any missing modules referenced (../feishu-common, ../evolver) to ensure they are honest and available; 4) Search the repo for all places that read files (memory/, logs/, assets/) to understand what might be uploaded to Feishu; 5) If you do not trust the source, do not provide production credentials and run only in a throwaway environment. The skill appears to implement its claimed features, but the omission of declared credentials and broad system access make it worth manual review before use.

Type: OpenClaw Skill Name: feishu-evolver-wrapper-local Version: 1.0.0 The skill bundle implements a complex 'self-evolution' loop with several high-risk behaviors that, while aligned with its stated purpose, create a significant attack surface. Key concerns include the use of 'new Function()' in 'index.js' to evaluate potentially malformed LLM output (creating an RCE vulnerability), the automatic installation of a persistence mechanism via OpenClaw's cron system in 'lifecycle.js', and an 'auto-heal' feature in 'skills_monitor.js' that executes 'npm install' on other skill directories. While the inclusion of a secret-scanning utility in 'feishu-helper.js' suggests a lack of malicious intent, the combination of broad shell execution, self-modifying git operations, and automated persistence warrants a suspicious classification.