The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill retrieves sensitive credentials using shell commands and database queries:
Accesses the ig_bot_config database table via psql to extract API keys: SELECT config->>'apiKey' FROM ig_bot_config.
Extracts a Redis password from a sensitive environment file: grep ^REDIS_PASSWORD= /opt/superseller-worker/apps/worker/.env.
[COMMAND_EXECUTION]: The skill makes extensive use of powerful shell commands that interact with system internals and remote infrastructure:
Uses ssh root@172.245.56.50 to perform operations on a remote server with root privileges.
Uses sed -i to modify the contents of the .env file in the application directory: /opt/superseller-worker/apps/worker/.env.
Executes docker exec to run commands inside a container using extracted passwords.
[REMOTE_CODE_EXECUTION]: The skill initiates remote code execution on an external server using npx tsx to run TypeScript scripts: ssh root@172.245.56.50 '... npx tsx src/scripts/schwed-photoday-v4.ts'.
[DATA_EXFILTRATION]: The skill performs sensitive file reads (.env files) and database queries (psql) combined with network operations (ssh, curl), creating a high-risk path for data exfiltration.

kie-byok-stuck