The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructs the user to install the Fabric tool using a 'curl | bash' one-liner: curl -fsSL https://raw.githubusercontent.com/danielmiessler/fabric/main/scripts/installer/install.sh | bash. This pattern executes remote code without prior inspection, bypassing local security controls.
[COMMAND_EXECUTION]: The skill requires the Bash tool to execute various CLI operations, including pattern updates (fabric -u), environment setup (fabric --setup), and custom pattern creation using heredocs (cat > ... << 'EOF').
[EXTERNAL_DOWNLOADS]: The skill's core functionality involves downloading and updating prompt 'Patterns' from a remote GitHub repository through the fabric -u command.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its processing of untrusted data:
Ingestion points: Data enters the context via YouTube video transcripts (fabric -y) and file reading (cat article.txt | fabric).
Boundary markers: The instructions do not specify any delimiters or safety warnings to prevent the agent from obeying instructions embedded within the processed text.
Capability inventory: The skill has access to Bash, Write, and Read tools, which could be exploited if an injected instruction targets the local system.
Sanitization: There is no evidence of sanitization or filtering applied to the external content before it is passed to the AI patterns.

fabric