The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes instructions to execute git and GitHub CLI commands to manage code changes and automate Pull Request creation. Evidence is found in references/pr-review.md, which details commands for pushing branches and invoking gh pr create.
[INDIRECT_PROMPT_INJECTION]: The skill processes findings from external reviewers and interpolates this content into generated descriptions and reports, creating a potential surface for indirect prompt injection. * Ingestion points: Native reviewer results and specialist deliveries from tools like react-development and platform-engineering (SKILL.md, references/stack-findings.md). * Boundary markers: No specific delimiters or boundary markers are specified to isolate external findings from the agent's context. * Capability inventory: Ability to write to the file system (e.g., .team-review/ directory) and execute shell commands (git, gh). * Sanitization: No evidence of sanitization or escaping of external findings is documented before they are included in reports or PR bodies.

review-consolidation