The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill reads and processes external, untrusted data to generate new repository content.
Ingestion points: The agent reads existing README files, commit history, and repository metadata using gh repo view and git commands.
Boundary markers: There are no explicit instructions or delimiters provided to the agent to distinguish between its instructions and the potentially adversarial content within the repository files.
Capability inventory: The skill has the ability to modify the repository by writing new files, committing changes via git, and updating GitHub metadata via gh repo edit.
Sanitization: There is no evidence of filtering or sanitizing the input data before it is interpolated into prompts for rewriting the description and topics.
[COMMAND_EXECUTION]: The skill executes several local commands to manage the repository and its assets.
It uses gh (GitHub CLI) to interact with the GitHub API for viewing and editing repository metadata.
It uses git for local repository operations such as tagging, committing, and pushing changes.
It runs a bundled Python script (render_terminal_gif.py) to process JSON storyboards and render GIF animations.
[EXTERNAL_DOWNLOADS]: The skill interacts with external services to verify assets.
It fetches badge images from Shields.io to ensure they are valid and return the correct status codes.
It uses curl to verify the presence of newly created assets on GitHub's raw content servers.

repo-showcase