The Agent Skills Directory

[PROMPT_INJECTION]: Indirect Prompt Injection surface detected. * Ingestion points: The skill reads commit messages using git log and git show (SKILL.md). * Boundary markers: There are no instructions to treat commit data as untrusted or to use delimiters to prevent the agent from obeying instructions embedded in the repository history. * Capability inventory: The skill has the ability to write HTML files to the local directory and execute shell commands including git and open. * Sanitization: Commit messages are distilled into a summary without explicit sanitization or validation steps.
[COMMAND_EXECUTION]: Local shell command execution for repository interaction and display. * Evidence: The skill executes git, wc, awk, and open commands to retrieve logs, calculate statistics, and display the generated HTML file. These operations are aligned with the skill's primary purpose.

changelog