The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses git commands (such as git log, git diff, and git rev-parse) to audit local copies of upstream repositories. These commands are executed within the scope of local paths defined in the user's upstream/sources.yaml configuration file and are used for read-only analysis during the audit phase.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data by reading commit messages and repository structures from upstream sources. This is a core functional requirement for an auditing tool. The risk is mitigated by the fact that the skill primarily generates informational Markdown summaries for the user and requires explicit human confirmation before performing any write operations to the upstream/ configuration directory.
Ingestion points: upstream/sources.yaml, last-sync.yaml, and the output of git commands executed on upstream local paths.
Boundary markers: None explicitly mentioned for processing git logs.
Capability inventory: Access to the local file system (specifically the upstream/ directory) and git command execution.
Sanitization: The skill relies on standard agent processing of text data to generate summaries.

custom-skills-upstream-ops