The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and act upon untrusted data from external sources, specifically comments on Meticulous test runs and check results. These comments are treated as the primary source of instructions for code changes.
Ingestion points: The agent retrieves external instructions via meticulous agent diff-comments and meticulous agent check-comments (or their MCP equivalents).
Boundary markers: There are no explicit boundary markers or instructions to disregard malicious directives embedded within the processed comments. The skill explicitly tells the agent to "just work out what needs to change and make it happen."
Capability inventory: The agent has the capability to modify the local source code, commit changes, push code via git, and post comments to pull request platforms using gh, glab, or twg tools.
Sanitization: The instructions do not define any sanitization or validation logic for the external comment content before it is interpreted by the agent.
[COMMAND_EXECUTION]: The skill extensively uses shell commands and CLI tools to interact with the development environment and external services. This includes the meticulous CLI for test data, git for version control, and platform-specific tools like gh (GitHub), glab (GitLab), and twg (Teamwork Graph) for interacting with pull requests.

meticulous-fix