The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands, specifically git log, to extract development history for automated changelog generation as detailed in references/shipped-changes.md.
[INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because it ingests and processes untrusted data from the repository's version control metadata. 1. Ingestion points: The agent reads pull request titles, descriptions, commit logs, and changed files which are potentially controlled by external contributors. 2. Boundary markers: The skill instructions do not provide explicit delimiters or instructions to ignore potential commands embedded within the git metadata. 3. Capability inventory: The agent is granted the ability to execute local shell commands (git log), read files from the project repository, and generate marketing drafts. 4. Sanitization: No sanitization, escaping, or filtering is applied to the metadata retrieved from the repository before it is passed to the language model for processing.

launch