The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill executes external shell and Python scripts (score.sh and rhythm.py) located in a sibling directory (slop-sense/scripts/). These scripts are not included in the provided source files, making their behavior unverifiable within the context of this skill package.
[COMMAND_EXECUTION]: The skill instructs the agent to use the Bash tool for network operations and script execution. The specific command curl -sL <url> | sed 's/<[^>]*>//g' is vulnerable to command injection if the user-supplied URL is not sanitized by the agent, as shell metacharacters could trigger arbitrary execution.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from URLs, files, and pasted text, creating a surface for indirect prompt injection. 1. Ingestion points: SKILL.md (URL fetching and file reading tools). 2. Boundary markers: No delimiters or warnings are specified for the data passed to analysis scripts. 3. Capability inventory: Bash tool used for networking, file access, and script execution. 4. Sanitization: Basic HTML tag removal via sed on URL content.
[DYNAMIC_EXECUTION]: The skill dynamically runs scripts from the host filesystem at runtime using bash and python3, which relies on the presence of external files whose content and integrity are not controlled by the skill itself.

slop-check