The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources including GitHub pull request descriptions, commit messages, and local project files to generate newsletter content.
Ingestion points: Data enters the agent context through the outputs of gh pr view, gh pr diff, git log, and git diff commands, as well as via direct reading of local markdown and codebase files (e.g., SKILL.md Phase 1).
Boundary markers: There are no explicit instructions or delimiters defined to separate untrusted content from the system instructions or to warn the agent to ignore embedded commands within the ingested text.
Capability inventory: The skill has capabilities to read local files, execute shell commands (gh, git), and write files to the local emails/ directory (e.g., SKILL.md Phase 5).
Sanitization: No explicit sanitization or filtering logic is documented for the data ingested from PRs or logs before it is used in the writing phase.
[COMMAND_EXECUTION]: The skill uses local command-line interfaces to gather necessary data for its tasks.
Evidence: It explicitly calls gh pr view, gh pr diff, git diff, and git log to extract information about code changes and project history (e.g., SKILL.md Phase 1). This is standard behavior for its intended use case.

newsletter