The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill requires a MISP API key and instance URL for live operations. While it correctly encourages using CLI arguments for these, the handling of such secrets is inherent to its primary function as an API client.
[EXTERNAL_DOWNLOADS]: The skill installs the official pymisp library from PyPI, which is the standard and recommended tool for interacting with MISP instances.
[COMMAND_EXECUTION]: The instructions and code include commands for installing dependencies (pip install pymisp) and executing an internal Python agent (python agent.py). These are standard operations for a development and security tool.
[INDIRECT_PROMPT_INJECTION]: As a threat intelligence tool, it ingests external data (IPs, domains, hashes, URLs) that may originate from untrusted or malicious sources. While the skill performs validation (e.g., hash length, private IP checks), processing this data represents an attack surface for instructions embedded in IOC metadata.

performing-threat-intelligence-sharing-with-misp