The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the GitHub CLI (gh) to perform repository operations. It dynamically constructs commands for issue discovery, creation, and metadata modification, such as setting issue types via PATCH requests. It also performs GraphQL mutations to synchronize effort and priority fields with GitHub Project boards.
[DATA_EXFILTRATION]: Content from local documentation (PRDs, specs, and PLAN.md) is transmitted to GitHub's infrastructure. This data transfer is the core feature of the skill, enabling the conversion of local plans into remote tracking tickets on a well-known service.
[PROMPT_INJECTION]: The skill exposes an attack surface for indirect prompt injection by processing untrusted data from the issue tracker and local files.
Ingestion points: Fetches and reads full bodies and comments of existing GitHub issues and local project specifications.
Boundary markers: The instructions do not define protective delimiters or specific 'ignore' rules for content ingested from these external sources.
Capability inventory: The skill possesses file write capabilities to the local .harness directory and authenticated write access to GitHub repositories.
Sanitization: There is no evidence of explicit sanitization or validation of the ingested text before it is processed to generate new tasks.

harness-issues