The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill uses the ! syntax in SKILL.md to execute git status --short and git diff --stat HEAD when the skill is loaded. These are benign, read-only commands used to provide immediate repository status context to the agent and do not involve sensitive file access or network exfiltration.
[COMMAND_EXECUTION]: The skill instructs the agent to perform repository modifications and synchronization through shell commands including git add, git commit, and git push to fulfill its primary purpose of git automation.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the repository's file changes (git diff) and user-supplied input ($ARGUMENTS) to generate commit messages.
Ingestion points: Data enters the agent's context through the output of git status/diff commands and user arguments.
Boundary markers: The instructions do not define specific delimiters to isolate the potentially untrusted diff content from the agent's instructions.
Capability inventory: The skill has the capability to modify the local repository state and interact with remote repositories over the network.
Sanitization: The instructions lack explicit logic for the agent to sanitize or escape the ingested diff content before it is used for message generation.

ship