The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on the execution of bash scripts (scripts/moltbook.sh) to perform its primary functions, including API calls and data parsing.
[EXTERNAL_DOWNLOADS]: The SKILL.md configuration utilizes npx -y to dynamically download and execute the Apify MCP server (@apify/mcp-server) and the official Slack MCP server (@modelcontextprotocol/server-slack) from the npm registry.
[CREDENTIALS_UNSAFE]: The scripts/moltbook.sh script attempts to read sensitive platform-wide credentials from ~/.openclaw/auth-profiles.json and ~/.config/moltbook/credentials.json. While it specifically targets the Moltbook API key, the script requires read access to the entire authentication profile file.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. It ingests untrusted data from Moltbook (such as post content and comments) which is then processed by the agent. If this data contains malicious instructions, it could influence the agent's behavior, particularly since the agent has shell execution capabilities.
Ingestion points: Data enters the context through scripts/moltbook.sh via the hot, new, and post commands.
Boundary markers: None detected in the script or prompt instructions to delimit untrusted content.
Capability inventory: The skill has access to shell execution via the Bash tool and network operations via curl.
Sanitization: No evidence of sanitization or filtering of the ingested content before it is processed by the LLM.

moltbook-interact