The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted data in the form of git diffs and staged file content, which could contain malicious instructions designed to influence the agent's behavior during the review process.
Ingestion points: The skill reads output from git diff --cached and inspects the content of staged files (SKILL.md).
Boundary markers: The instructions do not specify explicit delimiters or warnings to separate the untrusted diff content from the agent's logic, although the task is strictly scoped to four specific categories of findings.
Capability inventory: The skill has the ability to execute shell commands (git, wc) and write its findings to the terminal.
Sanitization: There is no explicit requirement for the agent to sanitize or escape file paths or content before processing them.
[COMMAND_EXECUTION]: The skill requires the execution of shell commands, including git diff, git cat-file, and wc, to perform its core functionality. These commands are necessary for the skill's purpose of reviewing a repository's staged changes and are used in a standard local context.

ship-gate