The Agent Skills Directory

[PROMPT_INJECTION]: The skill ingests untrusted data from the repository's environment (indirect prompt injection surface).
Ingestion points: SKILL.md dynamically includes output from git status, git diff, and git log using the !command syntax.
Boundary markers: No explicit delimiters or instructions are provided to ignore potential commands embedded in existing commit messages or file diffs.
Capability inventory: The skill allows the agent to execute shell commands via git add and git commit.
Sanitization: No sanitization or escaping is performed on data ingested from git logs or diffs before it is presented to the agent context.
[COMMAND_EXECUTION]: The skill utilizes platform-specific syntax to execute shell commands at load time for context gathering.
Evidence: Commands such as !git status --porcelain, !git branch --show-current, and !git log --oneline -5 are executed automatically when the skill is loaded.
Analysis: These commands are used for local repository inspection and do not involve network exfiltration, sensitive file access outside the repository, or the injection of user-supplied arguments into the shell environment during the pre-execution phase.

ado-commit