The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from repository files such as commit messages, package manifests, and changelogs which could potentially contain prompt injection attacks.
Ingestion points: package.json, lockfiles, Git logs, and maintainer-contract.md.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are specified.
Capability inventory: The skill performs file reads, directory access (e.g., D:\Project\comet-website-docs), and command execution (build, test, git commands).
Sanitization: No sanitization or validation of the processed data is mentioned.
[COMMAND_EXECUTION]: The skill is designed to execute various development tools as part of the verification process.
Evidence: Instructions include running build, package dry-run, test suites, and Mintlify commands to verify release readiness.

comet-release