The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface which could be exploited by a malicious contributor via branch changes.
Ingestion points: The skill reads the output of git diff <base>...HEAD, which contains arbitrary code changes provided by users or external contributors.
Boundary markers: The instructions pass the raw diff content to sub-agents ("The full diff from Step 1") without using explicit delimiters or instructions to disregard embedded commands.
Capability inventory: The skill possesses significant capabilities, including the Edit tool (Step 4) to modify the codebase and shell access (Step 8) to execute git commit and git push to remote repositories.
Sanitization: No evidence of sanitization, escaping, or validation of the diff content is present before it is interpolated into sub-agent prompts.
[COMMAND_EXECUTION]: The skill performs various shell operations to inspect the project environment and manage the Git lifecycle.
Evidence: Commands used include git rev-parse, git diff, cat package.json, jq, rm -f /tmp/quality-gate-findings-*.md, git add, git commit, and git push.
[DATA_EXFILTRATION]: The skill includes functionality to transmit modified source code to a remote server via the git push command. While intended for standard development workflows, this constitutes a network operation that transfers repository data externally.

quality-gate