The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes git logs and commit messages from the repository environment. While this is necessary for its core functionality of generating changelogs and determining version bumps, it introduces a potential surface for indirect prompt injection if malicious instructions are embedded in commit history. * Ingestion points: Git logs (via stdin or text files) and release plan JSON files in changelog_generator.py and version_bumper.py. * Boundary markers: The instructions do not define specific delimiters to isolate commit messages from agent processing. * Capability inventory: Scripts provided use standard libraries only; examples in documentation suggest integration with network and shell tools. * Sanitization: Commit messages are parsed via regular expressions to extract structured metadata.
[EXTERNAL_DOWNLOADS]: The documentation and example scripts reference external services including GitHub API and DataDog for release planning and monitoring notifications. * Evidence: Examples in README.md and hotfix-procedures.md use the requests library to interact with api.github.com and api.datadog.com. These are well-known services and their inclusion follows industry standard practices for DevOps automation.
[COMMAND_EXECUTION]: The skill documentation provides numerous examples of shell commands for version control and deployment management. * Evidence: Scripts like release_planner.py include command strings for kubectl, and the README.md provides examples using git, npm, and docker. These are descriptive examples for user reference and are not executed automatically by the provided Python scripts.

release-manager