The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The script scripts/wespy_cli.py is designed to automatically clone an external repository (https://github.com/tianchangNorth/WeSpy.git) and immediately execute its main entry point via a dynamic import. This execution happens without any integrity verification, such as checking a specific commit hash or signature, which could allow for the execution of arbitrary malicious code if the upstream repository is compromised.
[COMMAND_EXECUTION]: The skill uses subprocess.run to execute shell commands for environment setup, specifically calling git clone to fetch the external dependency.
[DYNAMIC_EXECUTION]: The script scripts/wespy_cli.py manipulates the Python search path at runtime using sys.path.insert(0, str(repo_dir)) to load and run the wespy module from a newly downloaded, untrusted directory.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and process data from arbitrary external URLs (web articles). This represents a vulnerability surface where malicious instructions embedded in a webpage could be processed by the agent.
Ingestion points: The url argument passed to scripts/wespy_cli.py (found in SKILL.md usage examples).
Boundary markers: None identified; instructions are passed directly to the fetcher tool.
Capability inventory: The skill has the ability to write files to the local system (article downloads) and perform network operations.
Sanitization: No sanitization or filtering of the fetched content is performed by the skill wrapper before the agent processes it.

wespy-fetcher