The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute shell commands using version control systems like Git and Sapling (sl).
Evidence: Mentions and provides examples for git log, git commit, sl commit, git commit --amend, and sl metaedit in SKILL.md to interact with the repository.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze repository diffs to generate commit messages. This creates a surface for indirect prompt injection where malicious comments or code within the diff could attempt to influence the agent's generated output or behavior.
Ingestion points: The agent is instructed to "Read the diff" as the primary source for generating content (SKILL.md).
Boundary markers: There are no explicit instructions to use delimiters or ignore embedded instructions within the diff content.
Capability inventory: The skill has the capability to execute git commit and sl commit using the generated content, as well as read repository history via git log.
Sanitization: There is no mention of sanitizing or escaping the content read from the diffs before it is used to generate commit messages.

commit-msg