The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill fetches metadata and instructional files from external repositories to guide its behavior, which introduces a potential vector for indirect prompt injection.
Ingestion points: The skill retrieves CONTRIBUTING.md, CODE_OF_CONDUCT.md, and .github/PULL_REQUEST_TEMPLATE.md from the upstream repository in SKILL.md (Step 2).
Boundary markers: The instructions do not specify the use of boundary markers or safety delimiters when interpolating this external content into the agent's context.
Capability inventory: The skill has the capability to perform file reads, repository forking, branch creation, commits, pushes, and opening pull requests using the gh and git tools.
Sanitization: There is no evidence of sanitization, filtering, or validation of the fetched remote content before it is used to generate commit messages and pull request descriptions.
[COMMAND_EXECUTION]: The skill heavily relies on the execution of shell commands to interact with the local environment and GitHub.
Evidence: SKILL.md documents a workflow involving multiple shell invocations of git and gh CLI tools (Steps 1, 2, 3, 5, 6, 7, and 8) to manage version control and repository interactions.

lov-gh-contribute